Freelancer profile translated to English.

Description

AI Engineer Python / FastAPI specializing in the design of high-performance APIs and automation through AI & LLM.

I help companies modernize their architectures, make their systems reliable, and accelerate their processes with backend and AI solutions designed for production.

Some recent results:

• Python/LLM pipeline automatically consolidating ICPE permits for the Ministry of Ecological Transition (500k installations monitored)

• Refactoring of a Back-for-Front to FastAPI at Boulanger: -60% on response times in production

• OCR + AI pipeline with +900% quality on targeted business cases

• Dozens of REST endpoints on GCP (Cloud Run, Pub/Sub, Cloud SQL) at ADEO

Stack: Python, FastAPI, Django, Flask, LLMs (OpenAI, Anthropic, Mistral, Gemini), OCR, Docker, CI/CD, GCP, Azure, React.

Industry field of expertise

Languages

French
Native or bilingual
English
Native or bilingual

Workplace preferences

Can work on-site

Lille (up to 50km), Paris (up to 50km)

Ministère de la Transition écologique et de la Cohésion des territoires
AI Engineer
PUBLIC SECTOR
January 2026 - Today (7 months)
Paris 17 Batignolles-Monceau, France
Within the DGPR: Design and deployment of OCAPI, a Python 3.12 pipeline that automatically consolidates prefectural decrees for classified installations (ICPE) into a single, up-to-date legal permit, based on Géorisques PDFs — a high-value regulatory problem for the approximately 500,000 ICPE installations monitored by the State and their inspectors.

Architected the end-to-end pipeline (operation detection, chunking, graph resolution, HTML rendering) and delivered ~50 merged pull requests in 3.5 months, from the first commit to GitHub Pages publication. Integrated and benchmarked 10+ LLMs across multiple providers (OpenAI GPT-4o/5, Anthropic Claude Sonnet/Opus, Mistral, Google Gemini) via a reproducible .xlsx evaluation infrastructure, with specialized prompt engineering on French legal vocabulary.

Industrialized quality: strict mypy, pre-commit, exact HTML snapshot testing, centralized configuration (pydantic-settings), status code system propagated across the entire operation graph for fine-grained error traceability. Direct liaison with end-users (ICPE inspectors) to qualify error cases and prioritize developments. Stack: Python 3.12, Pydantic, BeautifulSoup, pytest, mypy, multi-LLM APIs, Git/GitHub Pages.
Python LLM Artificial Intelligence Automation API
Boulanger.com
AI Engineer
E-COMMERCE
April 2025 - December 2025 (8 months)
Lille, France
Complete refactoring of a legacy Back-for-Front to a Python / FastAPI REST API foundation:
redefinition of API contracts, simplification of workflows, optimization of data access, and implementation of automated tests, resulting in a 60% reduction in response times in production.
Integration of artificial intelligence components for data and document processing (OCR): design of an AI pipeline (OCR engine + post-processing) and optimization of models, leading to a 900% improvement in OCR results for targeted business cases and a significant increase in the quality of exploited data.
Python/Django Python Flask External API Integration
ADEO
Application Engineer
RETAIL (SMALL BUSINESS)
January 2024 - January 2025 (1 year)
Lille, France
Design and development of several dozen REST API endpoints (creation, evolution, and maintenance). Back-end development in Python / FastAPI within a GCP environment (Cloud Run, Cloud Functions, Pub/Sub, Cloud SQL), with the implementation of CI/CD pipelines and applications handling thousands of requests per day. Agile team collaboration (Scrum) on 2-week sprints: active participation in ceremonies, regular delivery of new features, PRs (2 to 5 user stories per sprint).
Python FastAPI Google Cloud SQL