Freelancer profile translated to English.
Back to original languageAbout Lucas
Independent AI Architect & Doctor (PhD) in Machine Learning.
I help CTOs move their Generative AI projects out of notebooks and into systems that truly run, without compromising data sovereignty or scalability. As an industrialization specialist, I design the technical backbone (Backend & Ops) essential in the era of Agentic AI.
🚀 WHY COLLABORATE WITH ME?
🔹 Pioneer of the MCP (Model Context Protocol) standard:
I break down data silos by implementing MCP servers. I create a universal interoperability layer enabling your LLMs to intelligently connect to your internal tools (CRM, ERP, DB) for autonomous and secure agentic workflows.
🔹 Sovereign & High-Performance LLMOps:
Expert in On-Premise deployments via vLLM on OpenShift (OCP) clusters. I guarantee complete confidentiality and minimal latency for regulated sectors, using robust asynchronous FastAPI backends.
🔹 ROI & FinOps Vision:
As a former co-founder, I integrate cost considerations from the design phase. I optimize your GPU resources to control your inference expenses while ensuring critical observability (OpenTelemetry, Grafana).
💼 REFERENCES & INDUSTRIAL IMPACT:
Banking & Insurance (Matmut): Architecture of sovereign GenAI solutions (Multi-Agent RAG) compliant with the AI Act.
New Space (SpaceAble): Scaled deployment of predictive pipelines for orbital safety (SSA).
AdTech (Criteo AI Labs): R&D on Transformers for ultra-low latency environments (NeurIPS Publication).
🛠 TECHNICAL STACK:
Agentic: MCP Servers, LangGraph, Pydantic AI, Multi-Agent RAG.
Production: vLLM, Ollama, Deep Learning Robustness, Hybrid Search.
Infrastructure: OpenShift (OCP), Kubernetes, Docker, GCP (Vertex AI), CI/CD.
🎯 MODALITIES:
Available for technical audits, AI roadmap design, or Lead AI Architect missions.
French
Native or bilingual
Can work on-site
Rouen (up to 50km), Paris (up to 50km)
Experience
- MatmutLead AI Consultant | Banking & Insurance IndustrializationBANKING AND INSURANCEMarch 2025 - April 2026 (1 year and 1 month)Rouen, FranceMission: Industrialization of critical Generative AI systems in the Insurance & Banking sector. Merging high-performance Backend engineering and LLMOps to deliver sovereign and secure infrastructures.Key Achievements:Agentic AI & Standards (MCP): Pioneer design and deployment of Model Context Protocol (MCP) servers. Creation of standardized interfaces to connect LLMs to siloed data, enabling the scaling of Agentic Workflows.Sovereign Infrastructure: Orchestration of On-Premise deployments (vLLM, Ollama) on OpenShift (OCP) clusters, ensuring absolute data confidentiality and compliance with regulatory requirements.Multi-Agent RAG (HIRAG): Design of hybrid search architectures (Vectors + BM25 via ElasticSearch & MongoDB) to maximize the contextual relevance of business responses.Performance & Observability: FinOps optimization of GPU/CPU resources on OCP. Implementation of industrial monitoring via OpenTelemetry & Grafana for tracking latency and model drift in production.Security by Design: Advanced API security via JWT/OAuth2 protocols and maintenance of an ultra-robust asynchronous code base (FastAPI, Pydantic).
- AI StartupCo-founder & Chief AI Architect | Agentic AI & Cloud StrategyTELECOMMUNICATIONSJune 2024 - February 2025 (8 months)Rouen, FranceTechnical and architectural leadership for several AI-native SaaS initiatives focused on Agentic AI. My role involved designing the entire technical backbone, from prediction model design to scalable cloud infrastructure.Key Achievements & Deep Tech Engineering:Agentic Architectures & RAG: Design and deployment of real-time prediction engines based on RAG (Retrieval-Augmented Generation). Advanced integration of Vertex AI Vector Search to ensure ultra-precise predictions based on heterogeneous data streams.Cloud-Native Infrastructure (GCP): Complete orchestration of the Google Cloud Platform ecosystem. Development of asynchronous data pipelines via Firebase Cloud Functions and deployment of frontier models via Vertex AI.Performance & FinOps: Drastic optimization of cloud inference costs through the implementation of prompt caching strategies and intelligent model arbitrage (Model Routing).Product Leadership: Rapid transition from research phase to an industrial MVP ready for the mobile market, ensuring high availability and minimal, secure latency for the end-user.
- SpaceAbleLead AI Scientist | Space Situational Awareness (SSA) & High-Stakes Predictive SystemsAVIATION AND AEROSPACENovember 2023 - June 2024 (7 months)Rouen, FranceScientific direction for the development of predictive ML products dedicated to space infrastructure security. Focus on transitioning research models to industrial inference pipelines on GCP.Key Achievements & Technical Impact:Trajectory Prediction & Safety (SSA): Design of advanced algorithms for Space Situational Awareness (SSA), including satellite trajectory prediction and real-time collision risk assessment (Conjunction Assessment).Scientific LLM Agents: Architecture and deployment of Scientific Chatbots (RAG Agents) capable of querying complex space data corpora to assist flight operators.Scalable Serverless Infrastructure: Deployment of highly scalable inference pipelines on Vertex AI, orchestrating the processing of massive telemetry streams with optimized latency.Space Weather Analysis: Implementation of predictive models for solar wind and space weather analysis, protecting the operational integrity of Low Earth Orbit (LEO) constellations.
Recommendations
Be the first to recommend Lucas
Help this freelancer shine by sharing your experience working together.
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
- PhD in Machine LearningINSA Rouen2023
- Summer School: Hi Paris, Machine LearningHEC School of Management2022Summer School: Hi Paris, Machine Learning
Certifications
- Object Oriented PythonUdemy2023