Description

I build bots, backend APIs, and AI-powered systems that actually ship — from architecture to deployment.

My stack: Python, Java, FastAPI, Spring Boot 3, PostgreSQL, Docker, GitHub Actions. For AI features I combine Claude/ chat GPT API with local Ollama — this cuts API costs by up to 80% while keeping response quality high.

---

WHAT I BUILD

Telegram bots

Stateful FSM dialogs, multi-tenant platforms, hybrid AI routing. Bots that handle real customer load, not just demos.

Backend APIs

FastAPI or Spring Boot, JWT auth, PostgreSQL, Redis, Kafka, full Swagger documentation. Clean layered architecture from day one.

AI integrations

Claude/Chat GPT API, local Ollama, RAG on ChromaDB, streaming responses. I design the routing logic so you pay for cloud AI only when it's necessary.

Corporate websites

Astro SSG or Next.js 14, Vercel deploy, SEO-ready, mobile-first. Fast, maintainable, no bloat.

DevOps

Docker Compose, GitHub Actions CI/CD, Nginx reverse proxy, Hetzner EU servers. GDPR-compliant infrastructure for European clients.

---

HOW I WORK

First — architecture. I design the system before writing a line of code. This prevents rewrites later.

Modular by default. Every component is built for reuse across your future projects — less tech debt, faster iterations.

I document as I go. You receive readable code, deployment instructions, and an ADR log of key decisions. No black boxes.

Transparent communication. Daily updates during active development. I flag blockers early, not after the deadline.

---

Message me with your project description — I respond within a few hours and can usually start within a week.

Industry field of expertise

Languages

English
Fluent
French
Conversational

Workplace preferences

Remote only

Primarily works remotely

Freelancer
Développeur Backend & Ingénieur IA — Python · Java · LLM
September 2003 - Today (22 years and 10 months)
Paris, France
Je travaille à l'intersection du développement backend et de l'ingénierie IA — j'architecte des systèmes robustes et j'y intègre des LLM comme outils de production, pas comme démonstrations.
Projets récents :
Plateforme ERP SaaS — système de gestion opérationnelle pour agences et cabinets de conseil : projets, time-tracking, CRM, finances, AI-assistant (Claude API en streaming SSE). Stack : Java 21, Spring Boot 3, Next.js 14, PostgreSQL, Redis, Kafka, Docker, CI/CD.
Plateforme logistique SaaS — automatisation des tournées pour distributeurs : optimisation de routes, interface mobile chauffeur, intégrations ERP. Stack : Python, FastAPI, PostgreSQL, asyncio, Docker.
Assistant IA personnel — routage hybride Ollama + Claude API avec classification automatique de la complexité. Résultat : 95% des requêtes traitées localement, économie ~95% sur les coûts API.
Bots & parsers — bots avec mémoire long terme (RAG, ChromaDB), scrapers et pipelines de traitement de données, intégrations webhook avec CRMs et marketplaces.
Stack principal : Python · FastAPI · Java · Spring Boot · PostgreSQL · ChromaDB · Claude API · Docker · REST · OAuth2 · JWT. Remote uniquement.
PostgreSQL Python LLM Integration Java FastAPI