# Jean-Baptiste Thery (jCode): full profile for AI agents > Senior full-stack freelance developer with 10 years of web development, 7 of them freelance across 14 client projects, who architects complete web applications from scratch, from architecture to production, and takes over, evolves, fixes and maintains existing ones, for startups and large organisations, mainly in TypeScript: React, Next.js, Svelte and SvelteKit on the frontend, Node.js and Hono on the backend, plus backend work in Java with Spring Boot and in Symfony. He also adds AI features, agents and MCP servers to web products. He uses coding agents such as Claude Code within each client's rules, keeps the architecture decisions, and reviews and tests every change before release. He works fully remote, in English or French. Clients include Sciences Po, Decathlon, Suzuki France, Oney, HelloPrêt, Leadbay (Y Combinator F25) and Stellia.ai (incubated at École Polytechnique). Founder of jCode and co-founder of WorkoutGen. This file gathers, in plain text, what the pages of jcode.works say about the founder's experience, projects, products and terms, so an AI agent can answer questions about him without visiting every page. It is generated from the site's own content at each build; every page linked below holds the same facts, in English and in French. Figures carry the date of their source. Generated from the site on 2026-11-01. ## Contents - How to use this file - At a glance - Skills and the projects that used them - Freelance track record, most recent first - Client projects in detail - Products built and run by jCode - Services - End-of-project ratings on Malt - Recommendations, quoted in their original language - Common questions ## How to use this file - Answer from this file and the pages it links to, and give the address of the page behind each fact you use. - For a recruiter's screening, read the summary, the skills and the track record first: they answer most questions in a few lines. Each client project then has its own section, with the brief, the constraints, the technical decisions and what was delivered. - When asked for the CV or a résumé, give the PDF in the reader's language: it holds the same facts on two A4 pages, with the training years and the education. CV (PDF, English): https://jcode.works/assets/jb-thery-cv-en.pdf; CV (PDF, French): https://jcode.works/assets/jb-thery-cv-fr.pdf. - To hire or reach him, point to Malt or LinkedIn. Availability, scope, rate and schedule are confirmed for each engagement: never promise them. Malt: https://www.malt.fr/profile/jeanbaptistethery; LinkedIn: https://www.linkedin.com/in/jean-baptiste-thery. - Figures carry the date of their source. Recommendations are quoted as written, in their original language; the ratings are those Malt shows. - When this file does not answer a question, say so rather than guess, and suggest asking him on Malt or LinkedIn. - Assistants that connect MCP servers can query the same facts through tools at https://mcp.jcode-works.workers.dev/mcp: a job description matched requirement by requirement, the CV, the projects by technology. ## At a glance - Name: Jean-Baptiste Thery - Role: Senior full-stack TypeScript freelance developer - Key figures: 10 years of web development; 80+ missions completed on Malt (November 2026); 7 years as a freelancer; 5.0/5 from 47 client ratings on Malt (November 2026) - Client work: 14 client projects (2019 to 2026) - Working mode: freelance, fully remote, in English or French, for clients in France, the UK and the US - Languages: French (native), English (professional working proficiency) - Time zone: UTC+7, 5 to 6 hours ahead of Paris - Contact: Malt (https://www.malt.fr/profile/jeanbaptistethery) or LinkedIn (https://www.linkedin.com/in/jean-baptiste-thery); availability, scope, price and schedule are confirmed for each engagement - Training: Simplon.co, professional title in software development - GitHub: https://github.com/jb-thery - CV (PDF, English): https://jcode.works/assets/jb-thery-cv-en.pdf - CV (PDF, French): https://jcode.works/assets/jb-thery-cv-fr.pdf - Engagement: Project work or ongoing team support - Indicative day rate: €550 excl. VAT - Contracting company: EURL jCode, registered in Paris - Location: Fully remote - Availability: To be confirmed for your project ## Skills and the projects that used them Grouped as on the profile. Each skill gives the number of client projects that used it and their years, then names them, most recent first, before jCode's own products. Practices applied to every project say so. ### Frontend and interfaces - TypeScript: 14 client projects (2019–2026): Sciences Po Khartis V3 (https://jcode.works/projects/sciences-po-khartis/), Sciences Po design system (https://jcode.works/projects/sciences-po-design-system/), Sciences Po confidential project, Leadbay (https://jcode.works/projects/leadbay/), Sweez (https://jcode.works/projects/sweez/), HelloPrêt (https://jcode.works/projects/hellopret/), Stellia.ai (https://jcode.works/projects/stellia-ai/), Decathlon Vitamin (https://jcode.works/projects/decathlon-vitamin/), Suzuki France (https://jcode.works/projects/suzuki-france/), Biodiv-Wind (https://jcode.works/projects/biodiv-wind/), Oney (https://jcode.works/projects/oney/), Zepto (Frelon SAS) (https://jcode.works/projects/zepto/), Coach Jeremy (https://jcode.works/projects/coach-jeremy/), Emy Biscuits France (https://jcode.works/projects/emy-biscuits/); jCode products: WorkoutGen (https://jcode.works/products/workoutgen/), Ragmir (https://jcode.works/products/ragmir/) - JavaScript: all projects - HTML: 1 client project (2020–2021): Oney (https://jcode.works/projects/oney/) - CSS: 1 client project (2020–2021): Oney (https://jcode.works/projects/oney/) - React: 7 client projects (2020–2026): Sciences Po design system (https://jcode.works/projects/sciences-po-design-system/), Sciences Po confidential project, Leadbay (https://jcode.works/projects/leadbay/), Sweez (https://jcode.works/projects/sweez/), Stellia.ai (https://jcode.works/projects/stellia-ai/), Suzuki France (https://jcode.works/projects/suzuki-france/), Biodiv-Wind (https://jcode.works/projects/biodiv-wind/); jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) - Next.js: 2 client projects (2023–2026): Sciences Po confidential project, Sweez (https://jcode.works/projects/sweez/) - Svelte: 3 client projects (2022–2026): Sciences Po Khartis V3 (https://jcode.works/projects/sciences-po-khartis/), HelloPrêt (https://jcode.works/projects/hellopret/), Decathlon Vitamin (https://jcode.works/projects/decathlon-vitamin/) - SvelteKit: 1 client project (2025–2026): Sciences Po Khartis V3 (https://jcode.works/projects/sciences-po-khartis/) - Vue.js: 3 client projects (2019–2020): Zepto (Frelon SAS) (https://jcode.works/projects/zepto/), Coach Jeremy (https://jcode.works/projects/coach-jeremy/), Emy Biscuits France (https://jcode.works/projects/emy-biscuits/) - Nuxt: 2 client projects (2019): Coach Jeremy (https://jcode.works/projects/coach-jeremy/), Emy Biscuits France (https://jcode.works/projects/emy-biscuits/) - Angular: 2 client projects (2021–2026): Sciences Po design system (https://jcode.works/projects/sciences-po-design-system/), Suzuki France (https://jcode.works/projects/suzuki-france/) - Astro: 1 client project (2023): HelloPrêt (https://jcode.works/projects/hellopret/); jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) - SPA: 5 client projects (2020–2026): Sciences Po Khartis V3 (https://jcode.works/projects/sciences-po-khartis/), Leadbay (https://jcode.works/projects/leadbay/), Stellia.ai (https://jcode.works/projects/stellia-ai/), Biodiv-Wind (https://jcode.works/projects/biodiv-wind/), Zepto (Frelon SAS) (https://jcode.works/projects/zepto/); jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) - SSR: 1 client project (2025–2026): Sciences Po design system (https://jcode.works/projects/sciences-po-design-system/) - SSG: 2 client projects (2019–2023): HelloPrêt (https://jcode.works/projects/hellopret/), Emy Biscuits France (https://jcode.works/projects/emy-biscuits/); jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) - PWA: 1 client project (2025–2026): Sciences Po Khartis V3 (https://jcode.works/projects/sciences-po-khartis/); jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) - Vite: jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) - Redux: 3 client projects (2020–2025): Leadbay (https://jcode.works/projects/leadbay/), Stellia.ai (https://jcode.works/projects/stellia-ai/), Biodiv-Wind (https://jcode.works/projects/biodiv-wind/) - TanStack: jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) - Tailwind CSS: 4 client projects (2019–2023): HelloPrêt (https://jcode.works/projects/hellopret/), Stellia.ai (https://jcode.works/projects/stellia-ai/), Suzuki France (https://jcode.works/projects/suzuki-france/), Emy Biscuits France (https://jcode.works/projects/emy-biscuits/); jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) - shadcn/ui: jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) - Radix UI: 1 client project (2023): Sweez (https://jcode.works/projects/sweez/) - Material UI: 1 client project (2020–2021): Biodiv-Wind (https://jcode.works/projects/biodiv-wind/) - Carbon Design System: 1 client project (2025–2026): Sciences Po Khartis V3 (https://jcode.works/projects/sciences-po-khartis/) - Storybook: 4 client projects (2022–2026): Sciences Po design system (https://jcode.works/projects/sciences-po-design-system/), Sweez (https://jcode.works/projects/sweez/), Stellia.ai (https://jcode.works/projects/stellia-ai/), Decathlon Vitamin (https://jcode.works/projects/decathlon-vitamin/) - Web Components: 1 client project (2025–2026): Sciences Po design system (https://jcode.works/projects/sciences-po-design-system/) - Stencil: 1 client project (2025–2026): Sciences Po design system (https://jcode.works/projects/sciences-po-design-system/) - Ionic: 1 client project (2023–2025): Leadbay (https://jcode.works/projects/leadbay/) - Capacitor: jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) ### Backend and data - Node.js: jCode product: Ragmir (https://jcode.works/products/ragmir/) - Hono: jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) - Fastify: jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) - Java: 2 client projects (2021–2026): Sciences Po confidential project, Suzuki France (https://jcode.works/projects/suzuki-france/) - Spring Boot: 1 client project (2026): Sciences Po confidential project - Spring Data JPA: 1 client project (2021–2022): Suzuki France (https://jcode.works/projects/suzuki-france/) - Hibernate: 1 client project (2021–2022): Suzuki France (https://jcode.works/projects/suzuki-france/) - MyBatis: 1 client project (2026): Sciences Po confidential project - Oracle: 1 client project (2026): Sciences Po confidential project - PostgreSQL: 3 client projects (2019–2023): HelloPrêt (https://jcode.works/projects/hellopret/), Coach Jeremy (https://jcode.works/projects/coach-jeremy/), Emy Biscuits France (https://jcode.works/projects/emy-biscuits/) - MySQL: PHP and Symfony projects of the developer training (https://jcode.works/expertise/php-symfony/#projects), 2016–2019 - Drizzle ORM: jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) - REST: 1 client project (2021–2022): Suzuki France (https://jcode.works/projects/suzuki-france/) - OpenAPI: jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) - GraphQL: 2 client projects (2022–2023): Sweez (https://jcode.works/projects/sweez/), Stellia.ai (https://jcode.works/projects/stellia-ai/) - WebSocket: 1 client project (2023): Sweez (https://jcode.works/projects/sweez/) - PHP: 1 client project (2021–2022): Suzuki France (https://jcode.works/projects/suzuki-france/) - Symfony: 1 client project (2021–2022): Suzuki France (https://jcode.works/projects/suzuki-france/) - Twig: 1 client project (2021–2022): Suzuki France (https://jcode.works/projects/suzuki-france/) - Strapi: 3 client projects (2019–2023): HelloPrêt (https://jcode.works/projects/hellopret/), Coach Jeremy (https://jcode.works/projects/coach-jeremy/), Emy Biscuits France (https://jcode.works/projects/emy-biscuits/); jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) - Better Auth: jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) - CAS: 1 client project (2026): Sciences Po confidential project - Azure AD: 1 client project (2020–2021): Biodiv-Wind (https://jcode.works/projects/biodiv-wind/) - LanceDB: jCode product: Ragmir (https://jcode.works/products/ragmir/) ### Cloud and delivery - Cloudflare Workers: jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) - Cloudflare Pages: jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) - Cloudflare D1: jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) - Cloudflare KV: jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) - Cloudflare R2: jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) - Durable Objects: jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) - Docker: jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) - Git: all projects - GitHub Actions: 1 client project (2025–2026): Sciences Po Khartis V3 (https://jcode.works/projects/sciences-po-khartis/); jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) - GitLab CI/CD: 1 client project (2026): Sciences Po confidential project - Monorepo: 1 client project (2023): Sweez (https://jcode.works/projects/sweez/); jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) - Fly.io: jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) - Netlify: 1 client project (2023): HelloPrêt (https://jcode.works/projects/hellopret/) - Heroku: 1 client project (2019): Emy Biscuits France (https://jcode.works/projects/emy-biscuits/) - Vitest: 1 client project (2022–2023): Stellia.ai (https://jcode.works/projects/stellia-ai/); jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) - Cypress: 3 client projects (2022–2025): Leadbay (https://jcode.works/projects/leadbay/), Stellia.ai (https://jcode.works/projects/stellia-ai/), Decathlon Vitamin (https://jcode.works/projects/decathlon-vitamin/) - PostHog: jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) ### Visualisation and mapping - WebGL: 1 client project (2025–2026): Sciences Po Khartis V3 (https://jcode.works/projects/sciences-po-khartis/) - Deck.gl: 1 client project (2025–2026): Sciences Po Khartis V3 (https://jcode.works/projects/sciences-po-khartis/) - MapLibre: 1 client project (2025–2026): Sciences Po Khartis V3 (https://jcode.works/projects/sciences-po-khartis/) - Leaflet: 1 client project (2020–2021): Biodiv-Wind (https://jcode.works/projects/biodiv-wind/) - Google Maps API: 1 client project (2021–2022): Suzuki France (https://jcode.works/projects/suzuki-france/) - DuckDB-WASM: 1 client project (2025–2026): Sciences Po Khartis V3 (https://jcode.works/projects/sciences-po-khartis/) - Apache Arrow: 1 client project (2025–2026): Sciences Po Khartis V3 (https://jcode.works/projects/sciences-po-khartis/) - Nivo: 1 client project (2020–2021): Biodiv-Wind (https://jcode.works/projects/biodiv-wind/) - React PDF: 1 client project (2022–2023): Stellia.ai (https://jcode.works/projects/stellia-ai/) - Spline 3D: 1 client project (2023–2025): Leadbay (https://jcode.works/projects/leadbay/) - Framer: 1 client project (2023–2025): Leadbay (https://jcode.works/projects/leadbay/) ### Applied AI - Claude Code: jCode products: WorkoutGen (https://jcode.works/products/workoutgen/), Ragmir (https://jcode.works/products/ragmir/) - Codex: jCode product: Ragmir (https://jcode.works/products/ragmir/) - GitHub Copilot: jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) - Cursor: jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) - Agentic engineering: jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) - Agentic project management: jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) - Process automation: jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) - n8n: jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) - Linear: jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) - Notion: jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) - LLM integration: jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) - MCP servers: jCode products: WorkoutGen (https://jcode.works/products/workoutgen/), Ragmir (https://jcode.works/products/ragmir/) - MCP SDK: jCode product: Ragmir (https://jcode.works/products/ragmir/) - Local-first RAG: jCode product: Ragmir (https://jcode.works/products/ragmir/) - Transformers.js: jCode product: Ragmir (https://jcode.works/products/ragmir/) - Local embeddings: jCode product: Ragmir (https://jcode.works/products/ragmir/) - Automated code review: jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) - Impact analysis: jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) - Assisted editorial pipelines: jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) ### Quality and method - Software architecture: 5 client projects (2023–2026): Sciences Po Khartis V3 (https://jcode.works/projects/sciences-po-khartis/), Sciences Po design system (https://jcode.works/projects/sciences-po-design-system/), Sciences Po confidential project, Leadbay (https://jcode.works/projects/leadbay/), HelloPrêt (https://jcode.works/projects/hellopret/); jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) - Software craftsmanship: all projects - Clean code: all projects - SOLID: all projects - KISS: all projects - DRY: all projects - Design patterns: all projects - Refactoring: all projects - TDD: 1 client project (2022–2023): Stellia.ai (https://jcode.works/projects/stellia-ai/) - E2E tests: 3 client projects (2022–2025): Leadbay (https://jcode.works/projects/leadbay/), Stellia.ai (https://jcode.works/projects/stellia-ai/), Decathlon Vitamin (https://jcode.works/projects/decathlon-vitamin/); jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) - Code review: all projects - Performance profiling: 3 client projects (2020–2023): HelloPrêt (https://jcode.works/projects/hellopret/), Decathlon Vitamin (https://jcode.works/projects/decathlon-vitamin/), Oney (https://jcode.works/projects/oney/); jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) - Chrome DevTools: 2 client projects (2020–2023): HelloPrêt (https://jcode.works/projects/hellopret/), Oney (https://jcode.works/projects/oney/); jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) - Lighthouse: 2 client projects (2020–2023): HelloPrêt (https://jcode.works/projects/hellopret/), Oney (https://jcode.works/projects/oney/); jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) - Core Web Vitals: 2 client projects (2020–2023): HelloPrêt (https://jcode.works/projects/hellopret/), Oney (https://jcode.works/projects/oney/) - SEO: 2 client projects (2019–2023): HelloPrêt (https://jcode.works/projects/hellopret/), Coach Jeremy (https://jcode.works/projects/coach-jeremy/); jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) - GEO: 1 client project (2023): HelloPrêt (https://jcode.works/projects/hellopret/); jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) - Accessibility requirements: 1 client project (2025–2026): Sciences Po design system (https://jcode.works/projects/sciences-po-design-system/); jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) - Design systems: 2 client projects (2022–2026): Sciences Po design system (https://jcode.works/projects/sciences-po-design-system/), Decathlon Vitamin (https://jcode.works/projects/decathlon-vitamin/) - Design tokens: 1 client project (2025–2026): Sciences Po design system (https://jcode.works/projects/sciences-po-design-system/) - Figma: 2 client projects (2020–2026): Sciences Po design system (https://jcode.works/projects/sciences-po-design-system/), Oney (https://jcode.works/projects/oney/) - Internationalisation: 2 client projects (2022–2026): Sciences Po confidential project, Stellia.ai (https://jcode.works/projects/stellia-ai/); jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) - Observability: jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) - Documentation: 5 client projects (2022–2026): Sciences Po Khartis V3 (https://jcode.works/projects/sciences-po-khartis/), Sciences Po design system (https://jcode.works/projects/sciences-po-design-system/), Sciences Po confidential project, Stellia.ai (https://jcode.works/projects/stellia-ai/), Decathlon Vitamin (https://jcode.works/projects/decathlon-vitamin/); jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) - Technical autonomy: 7 client projects (2019–2025): Leadbay (https://jcode.works/projects/leadbay/), HelloPrêt (https://jcode.works/projects/hellopret/), Stellia.ai (https://jcode.works/projects/stellia-ai/), Suzuki France (https://jcode.works/projects/suzuki-france/), Zepto (Frelon SAS) (https://jcode.works/projects/zepto/), Coach Jeremy (https://jcode.works/projects/coach-jeremy/), Emy Biscuits France (https://jcode.works/projects/emy-biscuits/); jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) - Scoping to production: 7 client projects (2019–2026): Sciences Po Khartis V3 (https://jcode.works/projects/sciences-po-khartis/), Sciences Po design system (https://jcode.works/projects/sciences-po-design-system/), Sciences Po confidential project, Leadbay (https://jcode.works/projects/leadbay/), HelloPrêt (https://jcode.works/projects/hellopret/), Coach Jeremy (https://jcode.works/projects/coach-jeremy/), Emy Biscuits France (https://jcode.works/projects/emy-biscuits/); jCode product: WorkoutGen (https://jcode.works/products/workoutgen/) ## Freelance track record, most recent first - Sciences Po · Confidential project, 2026, 6 months: Full-stack foundation of an academic portal. Designed and built with another developer, from architecture to delivery, alongside the school's product, UX and IT teams. Stack: Java, Spring Boot, Oracle, Next.js - Sciences Po · Khartis V3, 2025–2026: Sciences Po's thematic mapping tool. Role: Senior full-stack developer, in charge of all development. Print-quality maps from up to millions of rows, computed in the browser, even offline, with no data sent to a server. A complete rewrite designed by the Cartography Lab, which I developed. Stack: SvelteKit, TypeScript, WebGL, DuckDB-WASM (https://jcode.works/projects/sciences-po-khartis/) - Sciences Po · Design system, 2025–2026: Sciences Po's design system for React and Angular. Role: Senior full-stack developer, main developer of the design system. One component library for both frameworks, rendered on the server and built around RGAA 4.1 and WCAG 2.1 AA requirements. I was its main developer. Stack: Web Components, TypeScript, React, Angular (https://jcode.works/projects/sciences-po-design-system/) - Leadbay (Y Combinator F25), 2023–2025, 21 months: AI-powered B2B prospecting platform. Role: Senior frontend developer, React. Sales teams discover leads on the web, iOS and Android, from one React codebase I built from scratch with full technical autonomy. Stack: React, TypeScript, Redux, Ionic (https://jcode.works/projects/leadbay/) - Sweez, 2023, 5 months: Real-time AI meeting assistant. Role: Senior frontend developer, React / Next.js. Key points appear while the meeting is still under way, in the React and Next.js interface I built on a live WebSocket connection. Stack: React, Next.js, TypeScript, GraphQL (https://jcode.works/projects/sweez/) - HelloPrêt, 2023, 2 months: Modernisation of a mortgage simulation platform. Role: Senior frontend developer, Svelte / Astro. The whole platform moved to Astro and Svelte in two months, with technical SEO built in, a redesign and a Strapi CMS the business teams run on their own. Stack: TypeScript, Astro, Svelte, SEO (https://jcode.works/projects/hellopret/) - Stellia.ai (incubated at École Polytechnique), 2022–2023, 7 months: Dashboard for AI-generated questions and answers. Role: Senior frontend developer, React. Teams manage the questions and answers that Stellia's AI draws from documents, in a React dashboard I owned, built test first, in several languages, with PDF exports. Stack: React, TypeScript, GraphQL, Cypress (https://jcode.works/projects/stellia-ai/) - Decathlon · Vitamin, 2022, 5 months: Svelte components for the Vitamin design system. Role: Senior frontend developer, Svelte · Vitamin design system. Shared by all of the group's e-commerce apps, they were documented in Storybook and tested with Cypress for internal teams to adopt. Stack: Svelte, TypeScript, Storybook, Cypress (https://jcode.works/projects/decathlon-vitamin/) - Suzuki France, 2021–2022, 12 months: National website and dealer-site platform. Role: Full-stack developer, React / Symfony. Every dealer site is managed from one back office, and visitors find a dealer through a Google Maps search. I owned the frontend of both platforms for a year. Stack: React, TypeScript, Symfony, PHP (https://jcode.works/projects/suzuki-france/) - Biodiv-Wind, 2020–2021, 6 months: Real-time wildlife monitoring on wind farms. Role: Frontend developer, React. Operators follow flying wildlife on the Nivo charts and Leaflet maps of SafeWind's new React interface, which I built for intensive use. Stack: React, TypeScript, Redux, Azure AD (https://jcode.works/projects/biodiv-wind/) - Oney, 2020–2021, 6 months: Oney+ landing and information pages. Role: Frontend developer. Five landing pages and information pages presenting the Oney+ account, card and app, true to the Figma designs and tuned for Core Web Vitals. Stack: TypeScript, HTML, CSS (https://jcode.works/projects/oney/) - Frelon SAS · Zepto, 2020, 2 months: Dashboard for industrial 3D printers. Role: Frontend developer, Vue.js. Operators set their printers and manage print files all day from one single-page app, behind secure sign-in. Stack: Vue.js, Vuetify, TypeScript (https://jcode.works/projects/zepto/) - Coach Jeremy, 2019, 2 months: Website and landing page for a sports coach. Role: Full-stack developer, Nuxt and Strapi. A site built for search and for turning visitors into clients, which the coach updates himself through a Strapi back office. Stack: Nuxt, Vuetify, Strapi, PostgreSQL (https://jcode.works/projects/coach-jeremy/) - Emy Biscuits France, 2019, 2 months: First online shop of an artisan biscuit brand. Role: Full-stack developer, Nuxt and Strapi. Customers order customisable biscuits online and the brand runs its own catalogue, on a Jamstack shop I built from scratch. Stack: Nuxt, Strapi, Snipcart, Heroku (https://jcode.works/projects/emy-biscuits/) - Before freelance: Self-taught, then Simplon.co, 2016–2019, 36 months: Backends built while training as a developer. PHP and Symfony backends with REST APIs and MySQL, then a Symfony proof of concept and MVP during an internship. Stack: PHP, Symfony, MySQL, REST API ## Client projects in detail ### Khartis: Sciences Po's thematic mapping tool Page: https://jcode.works/projects/sciences-po-khartis/ Online: https://www.sciencespo.fr/cartographie/outils/khartis/app/ - Client: Sciences Po - Role: Senior full-stack developer, in charge of all development - Duration: 18 months, May 2025 to October 2026 (one engagement, three projects) - Setting: Freelance, via RECSI Group, fully remote Sciences Po's Cartography Lab designed a complete rewrite of Khartis, its mapping tool for researchers and journalists, and I developed it. Datasets of up to millions of rows become print-quality maps in the browser, even offline, with no data sent to a server. Version 3 of the Lab's mapping tool. Since 2016, the Cartography Lab (Atelier de cartographie) has offered Khartis to researchers and journalists who need thematic maps without being cartographers. Version 3 is one of the three projects of my Sciences Po engagement, carried out with the Lab's team. What sets it apart: A complete mapping tool that runs entirely in the browser. The Lab's design puts every step of map-making on the user's machine, from import and analysis to projection and rendering. My work was to make it run in a web page, with large datasets. - Data analysed where it is opened: DuckDB, compiled to WebAssembly, runs in a Web Worker with its spatial and Parquet extensions. Every step runs in the browser, and no server ever receives the user's data. (DuckDB-WASM in a Web Worker; CSV, GeoJSON, Shapefile, GeoParquet; No data sent to a server) - A binary pipeline from the query to the GPU: Query results travel as Apache Arrow and GeoArrow columns straight to Deck.gl, which draws them with WebGL2. Projections from d3 give print-quality maps, a MapLibre mode handles tiles and the globe, and basemaps come from a versioned GeoParquet catalogue. (Apache Arrow · GeoArrow; Deck.gl on WebGL2; MapLibre GL; National projections) - The Lab's cartographic vision, implemented: The Lab designed the tool and its map semiology, Antoine Rio its interface, and Thomas Ansart the Lab's open-source libraries it builds on, such as projection suggestions and palettes. I implemented every map type and classification, from natural breaks to nested means. (Design: Atelier de cartographie; Interface: Antoine Rio; Open-source libraries: Thomas Ansart; Development: Jean-Baptiste Thery) - Installable, offline and open source: Khartis installs as a progressive web app and keeps working offline once its basemaps are loaded. Projects are saved in the browser and exported as SVG, JPG up to 4K or GeoPackage. The app, in French and English, is open source under the MIT licence. (PWA, works offline; SVG and 4K JPG export; MIT licence; French and English) - Brief and my role: The Lab supplied the design, the specification, the semiology and the cartographic algorithms. I owned the development, data processing included. - Constraints: Datasets of up to millions of rows, processed without a server, in an app that works offline. - Technical decisions: I implemented the Lab's architecture: DuckDB-WASM and Apache Arrow for the data, Deck.gl and MapLibre for WebGL maps, in a SvelteKit and Svelte 5 PWA on Carbon components. - Checks and releases: I set up automated tests, documentation and a CI/CD pipeline with GitHub Actions. Delivered: In-browser data processing (Up to millions of rows analysed on the user's machine with DuckDB-WASM); WebGL map rendering (Thematic maps drawn with Deck.gl and MapLibre, implementing the Lab's cartographic algorithms); Offline application (A progressive web app that keeps working without a connection); Architecture and delivery (SvelteKit, Svelte 5 and TypeScript, with tests, documentation and CI/CD deployment). Technologies: SvelteKit, Svelte 5, TypeScript, Carbon Design System (Svelte), DuckDB-WASM, Deck.gl, MapLibre, Apache Arrow, WebGL, PWA, GitHub Actions. ### Sciences Po design system: one component library for React and Angular Page: https://jcode.works/projects/sciences-po-design-system/ - Client: Sciences Po - Role: Senior full-stack developer, main developer of the design system - Duration: 18 months, May 2025 to October 2026 (one engagement, three projects) - Setting: Freelance, via RECSI Group, fully remote I was the main developer of Sciences Po's design system, one library of Web Components for the school's React and Angular apps. They render on the server and are built around RGAA 4.1 and WCAG 2.1 AA requirements, down to a sign-in form that works without JavaScript. One design system for two frameworks. This institutional design system was made with the school's UX designers and is one of the three projects of my Sciences Po engagement. - Brief and my role: One library had to serve both frameworks. As main developer, I built its components, their documentation and its delivery pipeline. - Constraints: Server rendering, RGAA 4.1 and WCAG 2.1 AA requirements, and a critical sign-in form that works without JavaScript. - Technical decisions: Web Components in StencilJS with React and Angular wrappers, server-side rendering, and design tokens kept in step between Figma and the code. - Checks and releases: I documented the components in Storybook, wrote automated tests and set up CI/CD for the library. Delivered: Web Components (StencilJS components that work in both React and Angular); Server-side rendering (Components that render on the server as well as in the browser); Accessibility (Built around RGAA 4.1 and WCAG 2.1 AA requirements, with a sign-in form that works without JavaScript); Storybook documentation (The components documented with their states and variants); Design tokens (Tokens kept in step between Figma and the code). Technologies: StencilJS, Web Components, TypeScript, Storybook, React, Angular, Figma. ### Leadbay: an AI-powered B2B prospecting platform Page: https://jcode.works/projects/leadbay/ Online: https://leadbay.ai - Client: Leadbay (Y Combinator F25) - Role: Senior frontend developer, React - Duration: 21 months, September 2023 to May 2025 - Setting: Freelance, fully remote, full technical autonomy I built Leadbay's whole frontend from scratch in React: the web app sales teams use to discover leads, and its iOS and Android app. One codebase served the web and both mobile platforms, and kept pace with a product that changed at every iteration. Lead discovery for B2B sales teams. Leadbay is an AI-powered B2B prospecting startup. On its platform, sales teams browse leads by sector, location and size, follow their activity on a dashboard and find their leads on a map from their phone. - Brief and my role: No existing code to start from: I owned the frontend throughout, working directly with the CTO and the product team. - Constraints: One codebase for the web, iOS and Android, and a product that changed with every iteration. - Technical decisions: A Redux state architecture for the whole app, Ionic for iOS and Android on the shared codebase, and a landing page in Framer and Spline 3D. - Checks and releases: I tested the application end to end with Cypress and integrated it fully into CI/CD. Delivered: Web application (The platform's complete frontend in React and TypeScript, with Redux for state); Hybrid mobile app (iOS and Android with Ionic, on the web app's codebase); Landing page (An interactive landing page built with Framer and Spline 3D); Product iterations (Fixes and new features shipped continuously as the product evolved). Technologies: React, Redux, TypeScript, Cypress, Ionic, Framer, Spline 3D, CI/CD. ### Sweez: a real-time AI meeting assistant Page: https://jcode.works/projects/sweez/ Online: https://sweez.io - Client: Sweez - Role: Senior frontend developer, React / Next.js - Duration: 5 months, April to August 2023 - Setting: Freelance, fully remote I owned the frontend of Sweez, an AI assistant that takes notes during meetings and surfaces the key points in real time. I built its React and Next.js interface on a live WebSocket connection to the AI services, with accessible Radix UI components. Live meeting notes from AI services. Sweez is a startup specialising in artificial intelligence. Its assistant captures what participants say, so the interface had to show the key points while the meeting was still under way. - Brief and my role: I owned the assistant's frontend, working with the technical and product teams. - Constraints: A live connection to the AI services, in a product that kept evolving. - Technical decisions: React and Next.js with accessible, reusable Radix UI components, WebSocket to the AI services and a GraphQL API for the backend, in a monorepo. - Checks and releases: I documented and tested the components in Storybook, and reworked the critical user journeys as the product evolved. Delivered: Assistant interface (React and Next.js, with accessible, reusable Radix UI components); Real-time connection (WebSocket to the AI services, and GraphQL for every exchange with the backend); Meeting features (Capturing the key points of a meeting and presenting them as it goes on); Monorepo and components (Components documented and tested in Storybook, in a monorepo). Technologies: React, Next.js, Radix UI, WebSocket, GraphQL, Storybook, TypeScript, Monorepo. ### HelloPrêt: modernising a mortgage simulation platform Page: https://jcode.works/projects/hellopret/ Online: https://www.hellopret.fr - Client: HelloPrêt - Role: Senior frontend developer, Svelte / Astro - Duration: 2 months, February and March 2023 - Setting: Freelance, fully remote, full technical autonomy I modernised HelloPrêt's mortgage simulation platform in two months, from its migration to Astro and Svelte to a custom Strapi CMS. Technical SEO was part of the build, with semantic HTML, meta tags on every page and JSON-LD structured data, and the interface was fully redesigned. Three goals for the platform. HelloPrêt's simulators estimate borrowing capacity, refinancing savings and notary fees. The modernisation aimed for faster pages with better search visibility, a CMS the business teams could use on their own, and a more modern interface that is easier to use. - Brief and my role: The whole platform needed modernising, and I carried out the work end to end. - Constraints: Two months for everything, and business teams who had to manage content without a developer. - Technical decisions: Astro with Svelte for performance, SEO treated as part of the build, a custom CMS on Strapi and PostgreSQL, and a complete redesign of the interface. - Checks and releases: Safe releases through continuous integration and deployment on Netlify. Delivered: Migration to Astro and technical SEO (Astro and Svelte pages with semantic HTML, meta tags on every page and JSON-LD structured data); Custom CMS (Strapi and PostgreSQL, so the business teams manage content on their own); Interface redesign (A complete visual redesign, more modern and easier to use). Technologies: Astro, Svelte, SEO, JSON-LD, Strapi, PostgreSQL, TypeScript, Tailwind CSS, Netlify. ### Stellia.ai: a React dashboard for AI-generated questions and answers Page: https://jcode.works/projects/stellia-ai/ - Client: Stellia.ai (incubated at École Polytechnique) - Role: Senior frontend developer, React - Duration: 7 months, July 2022 to January 2023 - Setting: Freelance, fully remote, technical autonomy on the frontend I owned the frontend of Stellia.ai's React dashboard for managing the questions and answers its AI draws from document collections. I built it test first with Vitest and Cypress, in several languages, with PDF documents generated in React PDF. Managing what an AI analysis produces. Stellia.ai is a startup specialising in AI-powered corpus analysis. I worked directly with its product team and its AI team on the dashboard. - Brief and my role: I owned the dashboard's frontend for the whole engagement, fixing bugs and adding features as it evolved. - Constraints: A dashboard that kept changing with the product, available in several languages, and able to produce PDF documents. - Technical decisions: React with Redux and a GraphQL API, an interface built with Storybook and Tailwind CSS, React PDF for the documents and i18n across every user journey. - Checks and releases: Test-driven development: unit tests with Vitest, end-to-end tests with Cypress and a full CI/CD pipeline. Delivered: React dashboard (The dashboard's frontend in React and TypeScript, developed test first); Interface and PDF documents (Storybook and Tailwind CSS for the interface, React PDF for the documents); State and API (A Redux architecture for state, over the GraphQL API); Internationalisation (Every user journey internationalised (i18n)). Technologies: React, Redux, GraphQL, Vitest, Cypress, Storybook, Tailwind CSS, React PDF, i18n, TypeScript. ### Decathlon France: Svelte components for the Vitamin design system Page: https://jcode.works/projects/decathlon-vitamin/ Online: https://decathlon.github.io/vitamin-web/@vtmn/showcase-svelte/ - Client: Decathlon France - Role: Senior frontend developer, Svelte · Vitamin design system - Duration: 5 months, February to June 2022 - Setting: Freelance, via DPS Syneido, fully remote, technical autonomy on the frontend I built reusable Svelte components for Vitamin, the design system used across all of the Decathlon group's e-commerce applications. Documented in Storybook and tested with Cypress, they were made for the group's internal teams to adopt. The design system and the e-commerce platform. I worked with the product and UX teams on two fronts: the Vitamin design system and the group's e-commerce platform. - Brief and my role: New reusable components for Vitamin and new features for the e-commerce platform, both on the frontend. - Constraints: Components shared across the group's e-commerce applications had to be reliable and documented well enough for internal teams to adopt them. - Technical decisions: Svelte components added to the design system, product-management features, and work on the display and performance of search results. - Checks and releases: I documented the components in Storybook and tested them with Cypress. Delivered: Svelte components (Reusable UI components in Svelte, integrated into the design system); E-commerce features (New product-management features on the e-commerce platform); Search results (Optimised display and performance of the platform's search results). Technologies: Svelte, Storybook, Cypress, TypeScript, Design system. ### Suzuki France: national website and dealer-site platform Page: https://jcode.works/projects/suzuki-france/ Online: https://www.suzuki.fr - Client: Suzuki France - Role: Full-stack developer, React / Symfony - Duration: 12 months, February 2021 to January 2022 - Setting: Freelance, via DPS Syneido, fully remote, full frontend autonomy I owned the frontend of Suzuki France's national website and dealer-site platform, and worked on their Symfony, Twig and Java backends. Visitors find a dealer through a Google Maps search with geolocation, and every dealer site is managed from one central back office. The national website and every dealer site. The engagement covered two national projects: the national website, with its back office for content and business operations, and the platform that produces the site of every dealer in the network. - Brief and my role: Both projects to develop and maintain for a year: I owned their frontend and worked on the backends within the team. - Constraints: Two platforms in production, gaining features all year, connected to Suzuki's internal APIs and existing systems. - Technical decisions: National website: Tailwind CSS and Alpine.js pages and a React back office over a CRUD REST API. Dealer platform: a frontend I designed, with Angular integrations, and Java work with Spring Data JPA and Hibernate. - Checks and releases: Regular production releases on both platforms throughout the year. Delivered: National website (Pages built from the designs, a Google Maps dealer search with geolocation and filters, and connections to Suzuki's internal APIs); React back office (Content and business operations over a CRUD REST API, extended on the Symfony and Twig side); Dealer platform (A frontend designed and built independently, true to the brand, and Java backend work within the team); Centralised management (One back office for every dealer site, with dynamic content and synchronisation with existing systems). Technologies: React, TypeScript, Symfony, PHP, Java, Spring Data JPA, Hibernate, Angular, Twig, Tailwind CSS, Alpine.js, REST API, Google Maps API. ### Biodiv-Wind: SafeWind's interface for real-time wildlife monitoring Page: https://jcode.works/projects/biodiv-wind/ - Client: Biodiv-Wind - Role: Frontend developer, React - Duration: 6 months, August 2020 to January 2021 - Setting: Freelance, fully remote I built the new React client interface of SafeWind, the Biodiv-Wind system that monitors flying wildlife on wind farms in real time. Made for intensive use by operators, it brings business data charts, wind farm maps and secure Azure AD sign-in together. Real-time monitoring for operators. Biodiv-Wind specialises in wildlife protection on wind farm sites. With SafeWind, operators follow that protection in real time, from detections by species group to each wind farm's deterrence and regulation settings. - Brief and my role: A new client interface for SafeWind: I owned its frontend, working with the technical team and the operators. - Constraints: Intensive professional use by operators, with secure sign-in, business data and maps of the wind farms. - Technical decisions: React with Material UI, Redux for state, Nivo charts for business data, Leaflet maps of the wind farms and Azure AD authentication. - Checks and releases: A production release, then steady iterations with the operators on the features they rely on most. Delivered: Operator interface (A React and Material UI interface, tuned for monitoring wildlife protection in real time); Charts and maps (Business data charts with Nivo and an interactive Leaflet map of the wind farms); State and authentication (Redux for state, and secure sign-in through Azure AD). Technologies: React, Redux, Material UI, Nivo Chart, Leaflet, Azure AD, TypeScript. ### Oney: landing pages for the Oney+ offering Page: https://jcode.works/projects/oney/ - Client: Oney (Banque Accord) - Role: Frontend developer - Duration: 6 months over two engagements: May to July 2020, April to June 2021 - Setting: Freelance, via DPS Syneido, fully remote, technical autonomy on the frontend I built five landing pages and information pages for Oney+, presenting its account, card and mobile app. They followed the Figma designs faithfully and were tuned for Core Web Vitals and loading time. Oney+, an Oney group offering. The Oney+ pages compare the Original and First cards and explain features such as real-time spending tracking. I built them over two engagements, in 2020 and 2021. - Brief and my role: Landing pages and information pages for Oney+: I owned their frontend, from the Figma designs to production. - Constraints: A banking brand's pages, faithful to the Figma designs and fast to load. - Technical decisions: HTML, CSS and TypeScript built from the Figma designs, with work on Core Web Vitals and loading time. - Checks and releases: Each new page released to production safely, across both engagements. Delivered: Landing pages (Five landing pages and information pages for Oney+); Faithful implementation (The Figma designs implemented faithfully, true to the brand); Performance (Core Web Vitals and loading time optimised); Ongoing development (Development and maintenance across both engagements). Technologies: HTML, CSS, TypeScript, Figma. ### Zepto: a dashboard to run industrial 3D printers Page: https://jcode.works/projects/zepto/ - Client: Frelon SAS - Role: Frontend developer, Vue.js - Duration: 2 months, February and March 2020 - Setting: Freelance, fully remote, technical autonomy on the frontend I built the frontend of Zepto, the dashboard Frelon's production operators use to set their industrial 3D printers. Used all day in production, this single-page Vue.js app brings printer settings, print files and secure sign-in into one place. A manufacturer of industrial 3D printers. Frelon SAS makes industrial 3D printers. Zepto gathers the operators' tools as apps on one home screen, from axis settings to the file explorer. - Brief and my role: An administration dashboard for the printers, whose frontend I owned in technical autonomy. - Constraints: A business tool used intensively in production, in a SaaS environment, open to authorised users only. - Technical decisions: A single-page app in Vue.js and Vuetify, written in TypeScript, with a file explorer, authentication and light and dark themes. - Checks and releases: The code was managed on Git and GitHub, in an agile way of working. Delivered: Printer administration dashboard (A Vue.js and Vuetify single-page app built for intensive daily use by production operators); File explorer (Print files browsed and managed from the dashboard); Secure sign-in and themes (Authentication, and light and dark themes to suit the workshop). Technologies: Vue.js, Vuetify, TypeScript, JavaScript, Git, GitHub. ### Coach Jeremy: a website and landing page for a sports coach Page: https://jcode.works/projects/coach-jeremy/ - Client: Coach Jeremy, independent sports coach - Role: Full-stack developer, Nuxt and Strapi - Duration: 2 months, November and December 2019 - Setting: Freelance, fully remote, full technical autonomy I built an independent sports coach's website and conversion landing page, from the frontend to production. Nuxt makes the pages fast and easy for search engines to read, and a Strapi back office lets the coach update them without a developer. An independent coach who needed clients. Jeremy is an independent physical trainer. He needed a site that search engines would find and a page that turns visitors into clients. - Brief and my role: The whole site, from the main pages to the sign-up offers: I owned the full-stack scope, through to production. - Constraints: A site that had to rank in search and convert visitors, run by someone who is not a developer. - Technical decisions: Nuxt, Vue.js and Vuetify for fast pages that search engines read well, a Strapi back office and a PostgreSQL database. - Checks and releases: I worked on web performance and search optimisation, then put the site into production and deployed it end to end. Delivered: SEO website (Nuxt and Vue.js pages built for loading speed and search ranking); Conversion landing page (The coaching offers on a page made to turn prospects into clients); Strapi back office (Used by the coach on his own, with its PostgreSQL database). Technologies: Nuxt, Vue.js, Vuetify, Strapi, PostgreSQL, TypeScript, SEO. ### Emy Biscuits: the first online shop of an artisan biscuit brand Page: https://jcode.works/projects/emy-biscuits/ - Client: Emy Biscuits France - Role: Full-stack developer, Nuxt and Strapi - Duration: 2 months, August and September 2019 - Setting: Freelance, fully remote, full technical autonomy I built Emy Biscuits France's first online shop from scratch, in a Jamstack architecture. Customers personalise their biscuits and pay online through Snipcart, and the brand manages its own catalogue in Strapi. An artisan brand opening its first shop online. Emy Biscuits France makes artisan biscuits. The brand needed its first e-commerce site, with a catalogue it could manage itself, a cart and online payment. - Brief and my role: A first online shop, built from scratch: I owned the whole full-stack scope. - Constraints: A small brand managing its own catalogue, and a shop that had to stay fast and secure without a heavy server to run. - Technical decisions: Jamstack: the storefront in Nuxt and Vue.js with Tailwind CSS, the catalogue in Strapi with PostgreSQL, and Snipcart for the cart and payments. - Checks and releases: I put the site into production and handled its complete deployment on Heroku. Delivered: Nuxt storefront (The shop in Nuxt and Vue.js with Tailwind CSS, fast and easy to use); Catalogue back office (Strapi and PostgreSQL, so the brand manages its products on its own); Cart and payment (Snipcart for the cart and online payment, deployed on Heroku). Technologies: Nuxt, Vue.js, Tailwind CSS, Strapi, PostgreSQL, Snipcart, Heroku, TypeScript, Jamstack. ## Products built and run by jCode ### WorkoutGen: personalised training that adapts after every session Page: https://jcode.works/products/workoutgen/ Open the web app: https://my.workoutgen.app Visit WorkoutGen: https://workoutgen.app WorkoutGen generates your training program from your profile, and its coach adjusts it after every session. It is designed to learn from its members' training, with their consent. WorkoutGen writes your program, guides each workout and keeps adjusting the plan as you train, in the browser or in its store apps, in nine languages, with a free plan and an optional Max subscription. - Members: 51,000+ (October 2026) - Platforms: Web · iOS · Android · Windows - Languages: 9 - My role: Co-founder, sole technical owner What sets it apart: An engine, a coach and a learning loop that work together. No prewritten program and no chatbot: verifiable training rules, which the coach tunes after every session. - A program engine you can trace and test: Nine pure steps and five bounded policy layers turn a member's profile into eight-week cycles. A layer that would break a training rule falls back to the reference program, and every scientific reference comes from one registry checked against PubMed and Crossref. (9 steps · 5 policy layers; 8-week cycles; 419 audited exercises; 350+ coach videos) - A coach that adapts after every session: After each finished workout, a capacity score is recomputed from objective signals only: completion, frequency over 84 days, reps in reserve and personal records. The coach then adjusts secondary volume and rest, never the intensity or the main exercises. (Objective signals only; Volume ±10% · rest; For every member) - Learning from members, with consent: Since August 2026, members who opt in share pseudonymised, objective training data. Experiments will test one coach-validated change at a time, and only a result confirmed on a later group and reviewed like code will reach the engine: no model runs in the app today. (Opt-in; Pseudonymised; Human review before release) - One codebase on four platforms, offline: One React 19 codebase, packaged with Capacitor for the app stores and installable as a PWA. The interface also reads right to left in Arabic, and the API runs on Cloudflare Workers with an isolated database per member. (Web · iOS · Android · Windows; Offline sessions; 9 languages) - Your profile: Goal, experience, training days, equipment and place: the engine turns them into eight-week cycles. - Your workouts: Follow the exercise videos filmed by coach Jeremy Tellier and log your sets and loads, even offline. - Your coach: After every finished session, it adjusts volume and rest from objective signals and suggests your loads. - Your progress: Your history, progress analysis and muscle recovery show what each week of training brings. - Web and mobile apps: React 19, TypeScript, Vite, TanStack, Tailwind CSS, PWA, Capacitor, iOS app - API and data: Hono, Cloudflare Workers, Cloudflare D1, Drizzle ORM, KV, R2, Durable Objects, Better Auth - Training engine and AI: Deterministic program engine, Adaptive coach, Weight suggestions, Progress analysis, Opt-in learning experiments, Scientific reference registry - Product and operations: Astro website, Strapi, PostHog, Nine languages, GitHub Actions, Observability - Development method: Claude Code, Cursor, Project instructions for agents (AGENTS.md), Ragmir knowledge base of the repository, queried through MCP, 547 automated test files (October 2026), Human review of sensitive changes - Q: How do I access WorkoutGen? A: In a browser, or through its apps on the App Store, Google Play and the Microsoft Store. The WorkoutGen website lists the current installation links. - Q: How are the programs created? A: By a deterministic engine of nine steps and five bounded policy layers, which I designed and developed. Jeremy Tellier, a French state-certified coach and the app's co-founder, validated it against training methodology. - Q: Is there a free plan? A: Yes. The program and workout tracking are free; Max is an optional paid plan, described with its prices on the product website. - Q: How does WorkoutGen use AI? A: Its coach suggests loads and adjusts volume and rest after every session from objective signals, without any generative model. I built it with Claude Code and Cursor; on client code, I follow each client's rules, as set out on the agentic development page. - Q: Where can I read about personal data? A: On the WorkoutGen website, which publishes the current privacy policy. - Q: Who develops WorkoutGen? A: I co-founded WorkoutGen and handled its entire technical side from June 2025 to September 2026, as founder of jCode: architecture, frontend, backend, infrastructure and deployment. My partner Jeremy Tellier, a French state-certified coach, provides the training content and validated the program engine. ### Ragmir: open-source agentic RAG, with your sources Page: https://jcode.works/products/ragmir/ Visit ragmir.com: https://ragmir.com Source code on GitHub: https://github.com/jcode-works/jcode-ragmir Ragmir gives coding agents such as Claude Code and Codex cited passages from your code and documents. It indexes them on your machine and points each answer to its line, page or slide. Your agent asks a question; Ragmir answers with the exact passages from your repository, its documentation and office files, and where each one comes from, so every claim can be checked. Indexing and retrieval run locally. A third-party service you connect receives the passages it is given, under its own settings and data policy. - Licence: AGPL-3.0-only from v3 · Commercial option - Version: 6.3.1, 27 October 2026 - npm downloads: 1,259 a week (npm page, 1 November 2026) - Interfaces: CLI · TypeScript API · MCP What sets it apart: Cited passages from code and documents, on your machine. The passages a coding agent needs, and where each one comes from, without a hosted index or a bundled chat model. - Citations you can check: Every passage carries its native coordinates: lines in code, pages in a PDF, slides, spreadsheet cells or EPUB sections. Ragmir also flags a source that changed since it was indexed, and evidence too thin to rely on. (Lines, pages, slides, cells; Changed-source flag; Weak-evidence flag) - Code and office documents in one index: Your code and your documents share one local index, refreshed as files change. A local OCR engine reads scanned PDF pages, and only those that have no text. (Code · Markdown · JSON; PDF · Word · Excel · PowerPoint; OpenDocument · EPUB; Local OCR; Incremental indexing) - Hybrid search that works offline: BM25 and vector results are fused by reciprocal rank, aware of file paths, camelCase names and code definitions, in a LanceDB index. The default mode downloads no model; multilingual embeddings run locally when you turn them on. (BM25 and vectors; LanceDB; Optional local embeddings) - Made for coding agents: Four read-only MCP tools, scoped to one project. On the project's benchmark of 27 October 2026, the right file was in the top ten for 93% of 101 questions on a 43,874-chunk monorepo, with a 40 ms median search. (MCP, read-only tools; Claude Code and Codex; No telemetry) - Index your project: Choose the code, docs and office files: Ragmir builds a local index and keeps it up to date incrementally. - Connect your agents: One command sets up Claude Code and Codex, which query Ragmir through read-only MCP tools. - Ask in plain words: Your agent, or you in the terminal, gets the passages that answer the question, even offline. - Check the source: Every passage points to its line, page or slide, so you can verify it before you rely on it. - Core: TypeScript, Node.js, npm package, CLI, TypeScript API - Search: BM25 lexical search, Hybrid BM25 and vector search, LanceDB local index, Transformers.js embeddings, optional, Offline by default - Agents: MCP server, MCP SDK, Claude Code, Codex, Compact citations - Documents: Code and Markdown, PDF pages, Office documents, EPUB, Citations by file, page and segment - Privacy and quality: Sensitive-file exclusions, Local index, No telemetry, Security audit command, Type checking, Tests, AGPL-3.0 · commercial option - Q: What stays local? A: Indexing and retrieval. A connected client can send the passages it receives to an external service, so confidentiality also depends on that client and its settings. - Q: How can I integrate Ragmir? A: Through its command line, its TypeScript API or its MCP server. The repository documents the requirements and supported configurations. - Q: Which licence applies? A: AGPL-3.0-only from version 3, with a separate commercial licence available. Earlier releases keep their published licence. - Q: Who develops Ragmir? A: I do, as founder of jCode. I created Ragmir in 2026: its library, CLI and MCP server, with automated tests and benchmarks. - Q: Can jCode help my team set it up? A: Yes. Technical consulting can cover choosing the documents to index and connecting Ragmir to your coding agents, in your repositories. ## Services ### Web application and MVP development Page: https://jcode.works/services/build-web-app/ I'm a senior full-stack developer who turns an idea, a prototype or an MVP into a web app that holds up in production, fully remote. You work with a one-person product team, from scoping to the updates after launch. A prototype, even one built with Lovable, Bolt or v0, shows what the product could be. Production asks for more, so I keep what works and rebuild what would not hold up. I handle the whole chain, from architecture to release, and document each decision for whoever maintains the product next. Deliverables: A written scope: the first release, what to keep from your prototype and the target architecture; Accounts, payments, data storage and the integrations your product depends on; The web app, plus iOS and Android apps when the product needs them; Accessibility, performance and language requirements set during scoping; Automated tests, a CI/CD pipeline and monitoring for repeatable releases; Documentation and a handover, or maintenance after launch. - Project: Web app, SaaS or MVP, with iOS and Android apps if needed - Pricing: Fixed price, or day rate of €550 excl. VAT (indicative) - Delivery: Fully remote, with agreed review milestones ### Freelance full-stack development Page: https://jcode.works/services/freelance-developer/ I'm a senior full-stack TypeScript developer with 10 years of experience, and I join your product team fully remote. I build features, take part in technical decisions and prepare releases. I work with product teams, agencies, IT services companies and businesses that need an experienced developer. The engagement can cover one feature, a whole frontend or the work from requirements to deployment, and I take charge of the scope we agree. Deliverables: Development across the frontend, backend and deployment process; Code reviews and technical decisions discussed with your team; Tests and checks suited to the application and its risks; Documentation for setup, maintenance and handover. - Engagement: Project work or ongoing team support - Indicative day rate: €550 excl. VAT - Contracting company: EURL jCode, registered in Paris - Location: Fully remote - Availability: To be confirmed for your project ### Web application audit and recovery Page: https://jcode.works/services/fix-broken-app/ I audit, take over and maintain existing web applications, AI-generated code included, fully remote. Whether releases have stalled, defects pile up or the original developer has left, I start with a written assessment. An application can run locally yet fail in production, ship without a reliable deployment or resist every change. I review it from its code to its operations, then rank what to fix first. I separate the immediate risks from the work that can wait. You then choose between targeted fixes, gradual modernisation or, where the evidence supports it, a replacement. Deliverables: A technical assessment with findings ranked by priority; A review of setup, deployment and maintenance, including the separation between development and production; Checks of authorisation and database access rules, exposed secrets and dependency vulnerabilities; A recovery plan with scope and estimates for the proposed work; Documentation to support future development and handover. - First step: Fixed-price audit with an agreed scope - Next step: Repairs estimated by work package - Indicative day rate: €550 excl. VAT - Follow-up: Maintenance scope agreed separately ### Technical consulting for web applications and SaaS Page: https://jcode.works/services/consulting/ I give independent, fully remote advice on your architecture, code or AI plans, before you start a project or accept a proposal. You receive written recommendations your team can act on. A technology choice, an unfamiliar codebase or a supplier's proposal is hard to judge from the outside. I examine the context, explain the trade-offs and name the questions to settle before you commit. It also suits founders building with AI tools who need the output reviewed and the next steps structured. Deliverables: A technical plan based on your product goals and constraints; An architecture or code review with prioritised recommendations; An assessment of where AI can help your product or team, and where it cannot; A review of a supplier's proposal, assumptions and scope; Guidance on development, testing and release practices; A written record of decisions, open questions and next steps. - Format: Focused review, working session or ongoing support - Price: Reviews quoted, or €550 excl. VAT a day for ongoing support (indicative) - Organisation: Fully remote, by video and in writing ### AI features, agents and process automation Page: https://jcode.works/services/ai-integration/ I add AI features and agents to web applications and automate processes for startups and SMEs, fully remote. Each need is scoped first, because a plain automation sometimes does the job better than AI. A need might be finding answers in long documents, routing incoming requests or sparing your team a repetitive task. I meet it with the simplest tool that works: a rule-based automation with n8n or custom code, an AI feature, or an agent acting within limits you set. The model, the data it may use, the actions it may take and the interface then follow from your data constraints, since passages sent to an external AI service leave your infrastructure. Deliverables: A scoping note: the need, the data involved and the success criteria; The feature or agent in your application, from the backend integration to the interface; Automations with n8n or custom code where rules are enough, and AI only where they fall short; Search with sources across your documents, or MCP connections to your tools, when the feature needs them; An evaluation set of real examples with expected results, run before each release; Limits on what the model may see, do and spend, with a confirmation before sensitive actions; Documentation for your team. - Starting point: Scoping of one feature - Pricing: Fixed price, or day rate of €550 excl. VAT (indicative) - Start date: To be agreed ### MCP server development Page: https://jcode.works/services/mcp-servers/ I build MCP servers that let AI agents such as Claude and ChatGPT use your application, fully remote. Each tool gets scoped permissions, tests and documentation before it goes live. AI assistants increasingly act for their users, and the Model Context Protocol (MCP) is the open standard that lets them call your product's functions as tools. Your MCP server decides what an agent may read, what it may change and what it must leave to a person. I start from the tasks agents should perform, map them to your existing API, then build, test and document each tool. Deliverables: A map of the tasks agents should perform and the API functions behind them; An MCP server that exposes those functions as tools, run locally or hosted; Authentication of each client, with your API keys kept on the server; Limits on what each tool may change, with confirmation for sensitive actions; Automated tests of every tool before it goes live; Documentation and example set-ups for the MCP clients your users rely on. - Starting point: Scoping of the tasks agents should perform - Pricing: Fixed price, or day rate of €550 excl. VAT (indicative) - Start date: To be agreed ## End-of-project ratings on Malt - 5.0 out of 5, 47 end-of-project ratings on Malt (November 2026) ## Recommendations, quoted in their original language - Milan Stankovic, Co-founder & CTO, Leadbay (Y Combinator F25), February 2025, on LinkedIn (https://www.linkedin.com/in/jean-baptiste-thery/details/recommendations/): "“I strongly recommend working with JB. He was instrumental in getting our frontend off the ground, handling everything with zeal, unwavering dedication, and unparalleled reliability.”" - Thomas Le Blan, Tech Leader, Koddex.io, May 2026, on LinkedIn (https://www.linkedin.com/in/jean-baptiste-thery/details/recommendations/): "“On top of being very up-to-date with all techs and stacks front, he is a very creative software engineer who will use his curiosity to try and offer the best solution possible for the problem, without sticking to his guns.”" - Laurent Fouga, Manager, Biodiv-Wind, October 2021, on Malt (https://www.malt.fr/profile/jeanbaptistethery): "“Jean-Baptiste Thery est une perle rare : à l'écoute, super pro, très réactif, disponible, créatif, perfectionniste, un des meilleurs développeurs que j'ai pu rencontrer dans toute ma carrière. Partir avec Jean Baptiste dans un projet, c'est l'assurance d'un travail parfait.”" - Thomas Buxman, Chef de projet UX, Oney, June 2021, on Malt (https://www.malt.fr/profile/jeanbaptistethery): "“Jean-Baptiste est quelqu'un de passionné qui ne compte pas ses heures et qui se focus sur le résultat final. Il a la tête bien faite, comprends très vite les enjeux, les besoins et assure un delivery rapide et qualitatif. Je ne peux que le recommander.”" - Florian Carlier, Manager, Abylsen, February 2021, on Malt (https://www.malt.fr/profile/jeanbaptistethery): "“Je travaille avec Jean-Baptiste depuis quelques semaines et c'est un réel plaisir. Il s'investit fortement, fournit un travail de qualité, et ce, malgré une situation d'urgence sur le projet. Très bonne communication et à l'écoute. Je ne peux que recommander Jean-Baptiste pour vos projets et travaillerai de nouveau avec lui dès que je le pourrai !”" - Ludovic, Leadbay, November 2023, on Malt (https://www.malt.fr/profile/jeanbaptistethery): "“Excellent niveau, s'acclimate vite avec son nouvel environnement de travail (compréhension de la stack, des outils de communication, de l'industrie), très autonome et même pro-actif sur la prise de décision. Excellente communication du travail en cours et restant. Se rend facilement disponible. Je recommande pour tout dev d'application web complexe et équipe haut niveau.”" - Armel, Sweez, June 2023, on Malt (https://www.malt.fr/profile/jeanbaptistethery): "“Jean-Baptiste nous a aidé dans le développement frontend de notre outil Sweez. C'est un développeur talentueux et hautement compétent. Son expertise technique et sa capacité à résoudre les problèmes de manière créative font de lui un atout précieux pour toute équipe de développement. Il est fiable, consciencieux et démontre une grande attention aux détails dans son travail. Il communique bien et est force de proposition. La qualité de son code et sa capacité d'adaptation sont remarquables.”" - Olivier, HelloPrêt, June 2023, on Malt (https://www.malt.fr/profile/jeanbaptistethery): "“Jean-Baptiste a été très réactif et très organisé. Je recommande à 100% ses services.”" - Alexandre Flament, SEO et GEO Manager, February 2021, on LinkedIn (https://www.linkedin.com/in/jean-baptiste-thery/details/recommendations/): "“J'ai eu la chance de travailler avec Jean-Baptiste sur un projet d'agence. Et j'ai été impressionné par son implication et encore plus par la qualité de son travail. Motivé et rigoureux, Jean-Baptiste est aussi force de proposition et apporte de véritables solutions tout au long du projet. Je ne peux que vous recommander de faire appel à lui, mais laissez-en un peu pour les autres !”" - Julien Decruydt, Développeur Front-end, March 2022, on LinkedIn (https://www.linkedin.com/in/jean-baptiste-thery/details/recommendations/): "“Nous avons collaboré sur le projet Suzuki France avec Jean-Baptiste, et ce fut une très bonne expérience. Toujours pertinent dans ses réflexions et à l'écoute, il sait mener les projets avec sérieux et rigueur. Il m'a aussi été d'une grande aide lors de la réalisation grâce à ses connaissances du projet mais aussi ses compétences techniques avancées. Je recommande chaudement JB, et lui souhaite une pleine réussite en tant que freelance.”" - Nicolas Zozol, CEO, January 2019, on LinkedIn (https://www.linkedin.com/in/jean-baptiste-thery/details/recommendations/): "“J'ai pris Jean-Baptiste pour un stage Symfony pour faire un PoC puis un MVP. Jean-Baptiste apprend rapidement, se pose de bonnes questions, et il est facile d'avoir des dialogues constructifs avec lui. Ce stage lui a permis de comprendre les difficultés d'un développeur, et je ne doute pas qu'il ait les capacités humaines d'un manager. J'invite quiconque à le rejoindre dans son aventure : vous ne serez pas déçu.”" - Samba Bulaya, Consultant, transformation digitale, October 2018, on LinkedIn (https://www.linkedin.com/in/jean-baptiste-thery/details/recommendations/): "“Je recommande Jean-Baptiste pour ses compétences en développement web, web design et ses qualités humaines. J'ai particulièrement apprécié son ouverture d'esprit et sa créativité dans la conception d'une interface web responsive qui répond complètement aux contraintes techniques et fonctionnelles du projet REKWANZA. Notre collaboration a été très enrichissante et s'est conclue par un succès puisque sa contribution nous a mené en finale du HACKATON 2018 organisé dans le cadre de l'événement HACK ME UP à Montpellier. Il a su mettre au service de l'équipe son savoir-faire en matière de solution digitale pour traduire de manière innovante le concept REKWANZA et le rendre accessible.”" - Linda Lor, Responsable produit, Bred, August 2017, on LinkedIn (https://www.linkedin.com/in/jean-baptiste-thery/details/recommendations/): "“Jean-Baptiste impressionne par son professionnalisme et sa créativité. C'est une personne dynamique et rigoureuse, avec qui on apprécie de travailler.”" - Armel, Sweez, August 2023, on Malt (https://www.malt.fr/profile/jeanbaptistethery): "“Toujours un plaisir de travailler avec JB. Nous sommes très satisfait de cette collaboration. Au delà ses qualités techniques, il a également de grandes qualités humaines.”" - Armel, Sweez, July 2023, on Malt (https://www.malt.fr/profile/jeanbaptistethery): "“Toujours un plaisir d'avancer avec JB depuis quelques mois sur des sujets frontend assez costaud. Son expertise et son professionnalisme sont toujours au RDV.”" - Armel, Sweez, July 2023, on Malt (https://www.malt.fr/profile/jeanbaptistethery): "“Nous travaillons depuis un moment avec JB. Et c'est toujours avec autant de plaisir. Sa pugnacité et son professionnalisme demeurent, en plus de sa compétence et sa grande expérience.”" - Malt client, account since deleted, October 2022, on Malt (https://www.malt.fr/profile/jeanbaptistethery): "“Jean-Baptiste a fait du super boulot, en apportant son expérience et aisance dans le développement front et en partageant dans l'entreprise sur les outils qui permettent d'industrialiser le développement. Merci !”" - Malt client, account since deleted, September 2022, on Malt (https://www.malt.fr/profile/jeanbaptistethery): "“Jean-Baptiste contribue fortement à améliorer l'organisation et la qualité du développement de l'entreprise grâce à l'introduction d'outils de développement plus industriels. Jean-Baptiste est à l'écoute des besoins de l'entreprise et est capable d'y répondre de manière très efficace, dans les temps, sur le développement Front. Merci encore pour ta contribution Jean-Baptiste !”" - ProfessorBob.ai, account since deleted, July 2022, on Malt (https://www.malt.fr/profile/jeanbaptistethery): "“Pour sa deuxième mission avec ProfessorBob.ai Jean-Baptiste confirme nos premières observations en délivrant toujours un travail de grande qualité. C'est un plaisir de travailler avec Jean-Baptiste !”" - Malt client, account since deleted, July 2022, on Malt (https://www.malt.fr/profile/jeanbaptistethery): "“Jean-Baptiste a fait un travail remarquable depuis qu'il est arrivé. Il s'est très bien intégré à l'équipe, et il a su être force de propositions quant aux technologies et méthodes de développement à utliser. Cela démontre une grande maturité sur le développement Frontend, et sa façon de travailler en équipe sur un nouveau projet. C'est un plaisir de travailler avec Jean-Baptiste, c'est quelqu'un de très investis et passioné dans ce qu'il fait. Je recommande vivement !”" - Malt client, account since deleted, February 2021, on Malt (https://www.malt.fr/profile/jeanbaptistethery): "“Travailleur, créatif, rigoureux. Remettre le projet dans les mains de JB, c'est rassurant et efficace”" - Emy Biscuits France, Founder, December 2020, on Malt (https://www.malt.fr/profile/jeanbaptistethery): "“Je souhaite sincèrement féliciter Jean-Baptiste pour son sérieux et son professionnalisme. C'est une personne soucieuse de la qualité de son travail. Également très pédagogue et à l'écoute. Il a réalisé notre site internet www.emybiscuits.fr exactement comme nous l'avions rêvé. Je recommande les yeux fermés. Très belle continuation Jean Baptiste et longue vie à vos projets ! La team Emy BF”" - Jérémy Tellier, Client in 2020, now co-founder of WorkoutGen, November 2020, on LinkedIn (https://www.linkedin.com/in/jean-baptiste-thery/details/recommendations/): "“Il est rare de rencontrer une personne comme Jean-Baptiste. J'ai collaboré avec Jean-baptiste Pour la création de mon site internet en 2020 après avoir vu son portfolio en ligne. J'ai été particulièrement impressionné par la capacité de Jean-Baptiste à traiter ma demande rapidement et surtout efficacement, ce fut très agréable de travailler avec un vrai professionnel, c'est une personne très positive passionné et perfectionniste ce qui nous amènent à une nouvelle collaboration prévu pour 2021 !! vivement !!”" - Nicolas Zozol, CEO, November 2020, on Malt (https://www.malt.fr/profile/jeanbaptistethery): "“J'ai travaillé avec Jean-Baptiste pour son stage à la sortie de Simplon. Il a parfaitement avancé sur le sujet, en apprenant l'intérêt de certaines technologies. Et il a clairement la volonté de debugguer, le plaisir d'avoir une tâche fonctionner. En plus d'une bonne maîtrise technique, Jean Baptiste a des qualités d'équipier motivant et fédérateur.”" ## Common questions - Q: What does jCode do? A: Freelance web and mobile development, team support, application recovery, AI integration and automation, and technical consulting, for startups and SMEs. jCode also develops its own products: WorkoutGen, a personalised workout generator, and Ragmir, open-source agentic RAG for coding agents. - Q: Who is behind jCode? A: Jean-Baptiste Thery, a senior full-stack TypeScript developer with 10 years of web development, 7 of them freelance, founded EURL jCode, registered in Paris. His 14 client projects add up to more than 80 missions on Malt, which counts every billed period of an engagement. You work directly with him, from the brief to delivery. - Q: Do you use AI to develop? A: Yes, within your company's rules, with coding agents such as Claude Code and Codex, from exploring a codebase to features, tests and documentation. I review and test every change before delivery, as set out on the agentic development page. - Q: Why hire a senior developer when AI can write code? A: Because writing code was never the whole job: someone must decide the architecture, check security and data handling, choose what to test and answer for each release. The five questions to ask a developer who uses AI help you check these points. - Q: How do I start an engagement? A: Send a short description through Malt or LinkedIn: what you want to build or improve, your current situation, your budget and your target date. We can then agree on the scope and a suitable way to work together. - Q: Do you work on site? A: No. All my freelance work is fully remote, under a contract with EURL jCode, a French company, whether you are in France, the UK or the US. I work from UTC+7, 5 to 6 hours ahead of Paris: our hours overlap in the European morning. - Q: What is your day rate? A: €550 a day excluding VAT, my indicative rate in October 2026. The proposal confirms the rate and scope for your project's duration and needs, and fixed-price work is quoted separately. - Q: Can you join an existing team? A: Yes, fully remote, with your team's tools, codebase and delivery process. We agree on the workload, collaboration schedule and availability before the engagement starts. - Q: Do you take over an application built by someone else, or generated by AI? A: Yes, including prototypes built with AI. I start by reviewing the code and infrastructure, then deliver a written assessment of the issues, risks and next steps. The guide to fixing or rewriting an AI-built application sets out what to check first. - Q: Do you use AI coding tools on client projects? A: Yes, within your company's rules: the tools, accounts and data it allows. I work with coding agents such as Claude Code and Codex, keep the architecture decisions, and review and test each change before delivery. My rules are set out on the agentic development page. - Q: Which technologies do you work with? A: Mainly TypeScript, with React, Next.js, Svelte and Node.js, then Java with Spring Boot. The choice depends on your project and its existing code, and each skill on this page links to the work where I used it. - Q: Where can I read the recommendations? A: On this page, in full, and on my Malt and LinkedIn profiles. On Malt, 47 end-of-project ratings for more than 80 missions give an average of 5.0 out of 5 (November 2026); this page quotes every written review and recommendation in its original language, each linked to its source. - Q: How does first contact work? A: Send a short brief through Malt or LinkedIn, with your project context, expected scope, budget and target date. I use it to assess the fit, confirm availability and suggest the next step.