Freelancer profile translated to English.
Back to original languageAbout David
I design and deploy AI systems in production for teams that want to move fast without breaking production: LLM agents with tool use, RAG pipelines on pgvector, multi-provider orchestration (Anthropic, OpenAI, Gemini) via LiteLLM, Langfuse observability.
Before AI, I spent several years architecting large-scale distributed data platforms — which allows me to deliver robust LLMs, not just demos.
French
Native or bilingual
Polish
Native or bilingual
Can work on-site
Paris (up to 20km)
Experience
- Chantal LängChantal's ScoresSOFTWARE PUBLISHINGJanuary 2022 - Today (4 years and 7 months)Genève, SwitzerlandLarge-scale crypto data aggregation platform (team of 4 developers).• Architected a Python data collection framework (42 sources, 13 blockchains, modular architecture) handling pagination, parallelism, and browser automation — mass MongoDB storage• Orchestrated 60+ distributed pipelines with Celery (hourly/daily/weekly scheduling) and asynchronous RabbitMQ publishing to decouple consumers• Implemented system resilience: fine-grained rate limiting, residential proxies, Pyppeteer browser automation for sources without public APIs, active memory monitoring on long-running workers• Developed the business API in NestJS/TypeScript (64 MongoDB aggregations, fundamental scoring, intelligent caching) with Prometheus + Grafana metrics and Docker deployment on AWSStack: Python · Celery · Redis · RabbitMQ · MongoDB · Docker · AWS · Pyppeteer · NestJS · TypeScript · Prometheus · Grafana
- IndépendantGolAiENTERTAINMENT AND LEISUREApril 2026 - May 2026 (1 month)Alfortville, FranceConversational AI assistant specialized in video games. Connected to the user's library (Steam, PSN), it answers questions about games, analyzes profiles, and provides personalized recommendations through semantic search.Architecture:• Pydantic AI agent with dynamic tool use and end-to-end SSE streaming• RAG pipeline: pgvector + sentence-transformers for semantic search on the user's library• Multi-provider LLM pipeline via LiteLLM (Anthropic, OpenAI, Gemini) with prompt caching (~60% API cost reduction) and Langfuse observability• Automatic game data enrichment pipeline (IGDB + SteamSpy + GPT-4o Mini Batch) generating 11 structured signals to refine RAG relevance• Extensible multi-platform import (Steam, PSN, designed to accommodate other stores without code modification)Stack: Python · FastAPI · React · PostgreSQL · pgvector · Pydantic AI · LiteLLM · Langfuse · sentence-transformers · Docker
- PersonnelHumantreeENTERTAINMENT AND LEISUREFebruary 2026 - April 2026 (2 months)Paris, FranceSolo design and development of a web platform for visualizing and managing skill trees, integrating a generative AI pipeline and a complete production infrastructure.
- Full async FastAPI REST API (SQLAlchemy 2.0, Alembic, Pydantic v2) with secure authentication via httpOnly cookies + JWT, and secret management via HashiCorp Vault
- Multi-provider LLM pipeline (OpenAI, Anthropic, Google Gemini) for automatic skill tree generation, with semantic search via pgvector
- PostgreSQL data model with complex many-to-many relationships
- React 19 / TypeScript frontend with interactive graph visualization (@xyflow/react + Dagre)
- Multi-stage Docker infrastructure + continuous deployment on Linux VPS via GitHub Actions (lint → test → build → deploy → smoke tests → automatic rollback)
- Complete observability: Prometheus/Grafana metrics, Loki structured logs, Tempo distributed traces, Discord alerting
- Integrated security: SAST Bandit, Trivy image scan, dependency audit, 40+ integration tests
🌐 In production: humantree.vps.webdock.cloud
Reviews
Recommendations
These freelancer profiles also match your criteria
Agatha Frydrych
Backend Java Software Engineer
4.7
(3)
2
Baptiste Duhen
Fullstack developer
4.6
(4)
5
Amed Hamou
Senior Lead Developer
4
(2)
7
Audrey Champion
Web developer
4.3
(3)
4
Education
- Apprenticeship engineering degree, Computer ScienceENSIIE - National School of Computer Science for Industry and Business2021Diplôme d'ingénieur en alternance, Informatique - Développement software (Java, Python, C, orienté objet, bas niveau, multithreading ...) - SGBD (SQL & NoSQL) - Réseaux (Configuration routeur, pare-feu, ...) - Docker - Développement web (PHP, NodeJS, Rest ...) - Sécurité (intro audit, SOC/SIEM, sécurité en environnement industriel) - Machine Learning - Data Science
- University technology diploma (DUT), Computer ScienceIUT of Montreuil2018Diplôme universitaire de technologie (DUT), Informatique