Marceau P.

Production AI Engineer | LLMs, RAG & Computer Vision

Production AI Engineer | Enterprise LLMs, RAG & Computer Vision ✓ Enterprise LLM systems with Python & AWS Bedrock ✓ Robotics, 3D Computer Vision & AI Research ✓ Master's in Mathematics, Vision & Learning (ENS Paris-Saclay) I help companies move AI from prototype to production. I'm a Production AI Engineer specializing in enterprise LLM systems, Retrieval-Augmented Generation (RAG), computer vision, and AI infrastructure. I work with startups, SaaS companies, and enterprise teams that need reliable AI systems integrated into real products—not just demos. What differentiates my work is the combination of advanced AI research and production engineering. My background spans quantitative modeling, computer vision, robotics, and enterprise AI, allowing me to bridge cutting-edge machine learning with scalable software systems that teams can deploy and maintain. I typically help companies: • Add AI capabilities to existing software products • Build production-ready LLM applications and RAG systems • Design scalable AI architectures • Solve technically challenging computer vision problems • Turn AI prototypes into reliable production systems Enterprise AI Systems I design and build: • Production RAG pipelines • AWS Bedrock and foundation model integrations • AI document intelligence platforms • Vector search and semantic retrieval systems • AI APIs and backend services • Evaluation, monitoring, and observability frameworks One of my recent enterprise clients wrote: "He built our LLM-powered features on Python and AWS Bedrock with rare depth of expertise, and he knows how to make AI systems reliable in production, not just functional." Beyond LLM systems, I also work on advanced computer vision and machine learning projects, including: • 3D point-cloud segmentation for robotics • Object detection and multi-object tracking • Semantic segmentation and perception systems • Research implementation and evaluation • Synthetic data generation • Model optimization and deployment For a robotics client, I built a proof of concept for semantic segmentation of 3D point clouds using NVIDIA's PartField implementation. My work focused on dataset preparation, 3D mesh curation, preprocessing pipelines, AWS experimentation, model evaluation, and adapting the workflow for robotic applications. I can support the complete AI lifecycle, including: • Technical architecture and feasibility studies • Data preparation and experimentation pipelines • Model and retrieval system development • Backend services and APIs • Cloud deployment and integration • Evaluation, monitoring, and documentation I hold a Master's degree in Mathematics, Vision, and Learning from ENS Paris-Saclay, one of Europe's leading AI programs. If you're building an AI product or integrating AI into an existing platform, I'd be happy to discuss your project and help design a solution that's ready for production.