Skip to content
Julien LeveneurFull-stack & AI engineer

About

I am a full-stack & AI engineer: I work with AI agents every day, and I also build AI pieces, such as MCP servers. I like taking a product from idea to server: modelling the data, writing the API, building the interface, then shipping it and keeping an eye on it.

I mostly work in TypeScript, with React and Next.js for interfaces and Node.js on the server. Agents write a good share of the code; my job is the frame that makes it reliable: the architecture, the repository's rules, and how a project is tested, deployed, and read again six months later.

This site is an example: a monorepo checked on every commit, a Next.js app deployed as a container on a VPS I run, and a contact form that goes through Resend.

How I work

  1. 01

    Typed end to end

    From the database schema to the component, through the API. No any, and every external input is validated at the boundary.

  2. 02

    Agents write, checks decide

    Reading code gives a hypothesis; a test gives evidence. An agent framed by types, lint, unit tests and journeys in a real browser ships fast without shipping just anything.

  3. 03

    Few dependencies

    Every dependency is a decision: pinned, justified, and removed as soon as it stops earning its place.

  4. 04

    Infrastructure is code

    Versioned with the application, replayed on every deploy, and reversible by redeploying an older commit.

Everyday tools

Front end
TypeScript · React · Next.js · Tailwind CSS · shadcn/ui
Back end
Node.js · Bun · tRPC · oRPC · BullMQ · NestJS · Better Auth
Data
PostgreSQL · Redis · Drizzle ORM · Prisma · Zod
Infrastructure
Docker · Kubernetes · Terraform · Caddy · GitHub Actions
AI agents
Claude Code · Codex · Cursor · MCP
Quality
Playwright · bun test · Oxlint · Knip · Turborepo