About Me

Hi, I'm Matthieu, an AI/ML Engineer currently developping the future @pigment. I am mainly interested in building and shipping production AI agents with node-based orchestration, parallel tool execution and solid eval/monitoring, taking them from prototype to reliable services.

I graduated from École Polytechnique and ENSAE Paris Tech with specialized focus in Artificial Intelligence, Natural Language Processing, and autonomous agent architectures.

For more detailed academic and professionnal experience, head to:

Current Endeavours & Projects

• 🚀 GenAI engineer - Core Agent Team - @pigment: Developping the low-level framework for AI agents builders at Pigment. Jungling with LangChain & LangGraph <3
• 🔬 Research Contributor - Co-authoring research paper under review for ICML 2026, contributing a new methodology for large datasets auto-augmentation
• 🎯 Focus Areas: MLOps, distributed systems, AI agent orchestration and deployment, ML models implementation and distribution

Achievements & Recognition

2025

• Datacraft (11/5/2025) - Presented the conference Enhancing Performance of RAG systems with mathematical methods in Paris
• IMC Trading International Challenge - Led Polytechnique team to top 1% global ranking (12,000+ teams)

2024

• Louis-Édouard Rivot Medal - Best Computer Science Research Internship (Polytechnique), awarded at Académie des sciences - Institut de France
• ICAIF 2024 - Crypto Market Simulation for Risk Estimation (built synthetic data generation systems for financial extreme scenarios simulation)
• MistralAI/HuggingFace × EntrepreneurFirst Hackathon - 4th place building a complete and optimized LLM web scraper for any-to-any api endpoint generation. Automatic web exploration, mapping and request retro-engineering.
• Digital Services (h)Acked Brussels - Implemented RAG-like systems for public-domain API documentation automation

Technical Expertise

ML/AI engineering: PyTorch advocate with WeightAndBiases/MLflow for training, Langchain & LangGraph for agentic development/orchestration + Langsmith for simple observability. LlamaIndex and LlamaParse for out-of-the box RAG, QDrant + PostgreSQL + Redis for complete semantic & agentic database system

DevOps & Infrastructure: Docker - Kubernetes - CI/CD with github/gitlab and CircleCI ( + Dokploy enthousiast) - AWS, OVH, GCP, Azure ML - Grafana , Prometheus, PgAdmin4/Adminer for monitoring.

Promising Open-Source Projects I Follow

• LangFlow: A cool all-in-one open-source tool for prototyping AI/RAG applications
• Immich: The best self-hosted photo manager yet, for your beautiful eyes
• Dokploy : An open-source alternative to vercel/Netlify for automatic deployment, previews and VPS cluster management.
• GoatCounter: A simple straightforward open-source web analytics tool for small projects

Let's Connect

• 📧 Email: contact@matthieu.olekhnovitch.fr
• 🌐 Portfolio: matthieu.olekhnovitch.fr