Description

PhD in Computer Science / Telco (IMT Atlantique). Technical co-founder with a decade of experience designing and shipping production AI systems: from ML pipelines at Sanofi to co-founding Namla.cloud, a Kubernetes-native, multi-tenant orchestration platform for cloud-and-edge workloads, with NVIDIA partnerships and SD-WAN integration. Prior to that, I worked several years as a low level Telco Research Engineer developing 4G/5G technical systems.

Core areas of expertise:

LLM systems & agent architecture: agent orchestration, multi-layer memory design, RAG and GraphRAG (cuGraph, NeMo Retriever, Nemotron Embed/Rerank)

Inference & performance engineering: Paged Attention, Flash Attention, KV cache optimization, speculative decoding, context-window strategy, cost-tiered model routing

Low-level systems: transformer internals (attention mechanics, KV cache layout), Linux systems and networking, performance profiling, kernel-level reasoning about throughput and latency

AI infrastructure & edge: Kubernetes, multi-tenant SaaS, GPU orchestration, edge fleets on Jetson, NVIDIA stack (NIMs, NemoClaw, Jetson, Brev)

Production stack: Python 3.12 / FastAPI, Postgres + pgvector, Next.js 15, SSE, APScheduler + Redis, Cloudflare Tunnel, VPS-to-scale architectures

How I work:

Comfortable as a Forward Deployed Engineer, Solutions Architect, or technical lead — equally at home architecting AI systems end-to-end, shipping production code, and translating between business stakeholders and deep technical teams. French/English bilingual, based in Paris, Open to remote engagements.

Domaines d’expertise

Langues

Français
Bilingue ou natif
Anglais
Bilingue ou natif
Arabe
Bilingue ou natif

Préférences en matière de lieu de travail

En télétravail uniquement

Travaille majoritairement à distance

Namla
Technical Co-founder
EDITION DE LOGICIELS
janvier 2022 - Aujourd'hui (4 ans et 5 mois)
Paris, France
Co-founded Namla.cloud and led the technical build-out of a Kubernetes-native, multi-tenant orchestration platform for cloud and edge workloads, with integrated SD-WAN networking and an NVIDIA partnership.

Defined the platform architecture end-to-end: multi-tenant control plane, distributed agent runtime, networking stack, and edge-to-cloud orchestration model.
Led a team of 4 engineers building the Namla Orchestrator backend (microservices architecture, Kubernetes operators, gRPC/REST APIs); owned engineering roadmap, technical hiring, and code review culture.
Owned the NVIDIA Jetson support layer — adapting the networking stack and agent runtime across multiple vendor hardware form factors and Jetson SKUs; shipped GPU-aware workload scheduling and remote lifecycle management for edge fleets.
Drove the technical relationship with NVIDIA Jetson and Metropolis teams (platform alignment, joint roadmap, co-marketing); platform architecture mentored by Sébastien Pahl (Docker co-founder), investor and board member.
Anchored technical credibility on strategic accounts: architecture reviews, deep-dive workshops, and on-site deployment with enterprise customers in telco, industrial, and defense.
Kubernetes Edge Computing Linux LLM Python
Sanofi Pasteur
Lead ML Engineer
INDUSTRIE PHARMACEUTIQUE
septembre 2020 - juin 2022 (1 an et 9 mois)
Rouen, France
Embedded ML engineering lead in pharmaceutical vaccine manufacturing — a regulated, high-stakes (GxP) environment requiring rigorous validation and production-grade reliability. End-to-end ownership of ML pipelines from raw industrial data through model training to production serving.

Designed and shipped production ML pipelines (Python, TensorFlow, OpenCV) for visual quality inspection and process optimization on the vaccine manufacturing line.
Applied NLP to virus sequence analytics for vaccine manufacturing optimization — processing large biological datasets to surface insights that directly informed production decisions.
Integrated ML workflows with AWS and enterprise data systems; operated autonomously across manufacturing, quality assurance, and data engineering stakeholders under GxP regulatory constraints.
Machine learning Python TensorFlow Computer Vision NLP
Mantu
Machine Learning Engineer
AGENCE & SSII
avril 2019 - septembre 2020 (1 an et 5 mois)
Nice, France
ML engineer in Mantu's innovation lab, owning the full ML lifecycle from training through production API serving for large-scale document intelligence.

Built an end-to-end NLP pipeline processing 100k+ resumes at scale: document ingestion, embedding generation, semantic matching (vector similarity search), and candidate–job relevancy scoring — deployed to production as RESTful inference services.
Designed asynchronous data pipelines (RabbitMQ) and a graph-based matching system (Neo4j) to power candidate–job recommendations.
Iterated rapidly with business stakeholders to align model outputs with real operational requirements.
Machine learning Python Deep Learning MongoDB Neo4j