AI Platform & GenAI Engineer
Designing self-hosted inference platforms, GenAI pipelines, and production systems that let organizations operate AI with more control, performance, and sovereignty.
Explore experienceI design and operate production-grade AI systems, from data pipelines to large-scale inference platforms.
My focus is turning AI into a controllable capability: running self-hosted models, building GenAI pipelines, and enabling organizations to produce and serve their own tokens instead of relying on external APIs.
I work across system design, distributed infrastructure, and model integration, with a strong emphasis on performance, sovereignty, and operational simplicity.
Skills
Experience
Education
Python, Kafka, Kubernetes, Docker, Terraform, CI/CD, distributed systems.
Inference platform design, sovereign AI architecture, performance optimization, observability, production operations.
Self-hosted LLM inference, RAG systems, LangGraph, vLLM, LiteLLM, WhisperX, model integration.
Excellent communication, problem solving, technical leadership, collaboration, fast adaptation.
Fluent French and English, intermediate Chinese and Japanese.
Designing AI training sessions for consultants, developers, and leadership around real-world AI systems and operations.
Leading sovereign inference platform initiatives so companies can become self-sufficient AI providers.
Leading the design and deployment of a sovereign inference platform for on-premise LLM, embedding, and speech serving.
Operating GenAI systems in production: RAG pipelines, LangGraph editorial tools, transcription, summarization, and metadata extraction.
Supporting recommendation products at media scale for 5 million+ users across a 100+ microservice architecture.
Stack: Python, Kafka, Neo4j, Kubernetes, Docker, Redis, AWS, Terraform, Ray, LiteLLM, vLLM.
Built internal data platform capabilities, including a configuration builder interface created from scratch.
Engineered DAG-based lineage querying, query validation, and deployable open-source RAG components.
Built a GenAI-augmented scraper for targeted collection, batch evaluation, and LLM-based classification.
Built dashboards and analytics workflows from delivery data to surface business insights.
Created a web application and ETL pipeline to automate scraping and streamline lead generation.
Stack: AWS, Docker, Dash, Streamlit, PostgreSQL, Selenium, Pytest, Pandas, SQLAlchemy.
Led the design and deployment of a complete data lifecycle: from web scraping and ingestion to transformation, storage, and dashboard-based analytics—empowering business decisions at Pragaz.
Built internal data platform capabilities at Amaris, from a configuration builder interface to DAG-driven lineage querying and validation workflows for more reliable platform operations.
Designed and deployed a local AI serving platform for LLM, embedding, and speech workloads, reducing external dependency while making GenAI usable inside production editorial systems.
Built production pipelines for transcription, summarization, metadata extraction, and RAG-driven editorial workflows to enrich content operations and improve discovery quality.
Maintaining and evolving a distributed recommendation architecture of 100+ microservices serving over 5 million users across news, VOD, podcasts, and live TV experiences.
Designed centralized logging for recommendation APIs and optimized Neo4j graph queries to improve response times, monitoring, and regulatory readiness.
Leading sovereign inference initiatives and AI literacy programs that help organizations understand, deploy, and operate their own AI capabilities instead of outsourcing them entirely to external APIs.
Made by Jean-Baptiste to show that I know a bit of HTML/CSS/JS