Summary
Mehdi Boubnan is a Senior AI Engineer specializing in building and scaling production AI platforms that bridge machine learning, backend engineering, and infrastructure. With nine years of experience and a strong applied-math background from CentraleSupélec and ENS Paris-Saclay, he has rebuilt ML serving stacks using ONNX, TensorRT, Triton and vLLM, and designed RAG and multimodal retrieval systems with pgvector. He routinely architects scalable backends (FastAPI, Celery, RabbitMQ, Redis, PostgreSQL), standardizes CI/CD and observability (Docker, ArgoCD, Prometheus, Grafana), and optimizes inference performance and GPU utilization. Beyond tooling, he focuses on developer ergonomics—dependency injection, lifecycle-managed services, and reusable data-access patterns—to reduce overhead and simplify operations at scale. Based in Paris, he combines research-grade model work (fine-tuning multimodal/visual LLMs) with production discipline and is open to senior platform and AI infrastructure roles across Europe, UAE, Morocco, and remote teams.
9 years of coding experience
7 years of employment as a software developer
(MVA) Master’s degree - Mathématiques Vision Apprentissage Applied Mathematics, (MVA) Master’s degree - Mathématiques Vision Apprentissage Applied Mathematics at ENS Paris-Saclay
Master’s Degree - MEng. Applied Mathematics, Master’s Degree - MEng. Applied Mathematics at CentraleSupélec
Classes préparatoires MPSI/MP, Classes préparatoires MPSI/MP at CPGE Moulay Youssef
French, Arabic, English