About Mohamed
- Development of real-time AI platforms (multi-agent, RAG, secure chatbots) deployed on-prem and cloud.
- Technical lead on strategic projects (Renault, Orange, Sephora, Indigo Retail), including computer vision, GNSS sensor fusion, and automated customer service.
- Architect and developer of several innovative generative AI-based solutions such as QuickResolve (multi-agent customer service), AI Coach (sports & nutrition), and an AI newsletter generation platform.
- Proven experience in DevOps & CI/CD (GitLab, Ansible, Observability via Prometheus/Grafana/Langfuse).
French
Native or bilingual
English
Native or bilingual
Arabic
Native or bilingual
Experience
- AXA Group OperationsAI Engineer LeadBANKING AND INSURANCENovember 2025 - Today (8 months)Paris, FranceDesign and deployment of security, evaluation, and governance solutions for generative AI and agentic applications, from defining the reference framework to multi-cloud production integration, including leading the internal technical community.
- Architecture of an enterprise GenAI gateway unifying AWS Bedrock, Azure OpenAI, and GCP Vertex AI behind a common layer of security, evaluation, and observability, serving as an organizational standard for LLM and agentic deployments.
- Automated evaluation framework for LLMs and agentic systems: adversarial test suites (prompt injection, indirect injection, jailbreaks, data exfiltration, memory poisoning) integrated into MLOps pipelines via Docker and CI/CD on Azure, AWS, and GCP.
- Vendor-neutral guardrails taxonomy (7 canonical controls: prompt injection, jailbreak, task adherence, PII detection, hallucination grounding, topic & scope, citation faithfulness) mapped to the native services of the three clouds and OWASP ASI Top 10.
- Secure multi-agent architecture: Guardrail patterns for LangGraph systems with MCP servers and A2A bus (Kafka/Redis); cloud coverage matrix and explicit mitigations for boundaries not covered by native services.
- Agentic governance: reference Agent Card schema for the internal AI agent registry; critical analysis of agent lifecycle (A2A/MCP, prompt drift, HITL vs HOTL).
- EU AI Act compliance: co-drafting AI governance standards with Security and Compliance teams; defining measurable model risk indicators.
- Leading the AI Engineering community: designing and leading the AgentOps workshop series (agent lifecycle, trigger taxonomy, nested Ops stack, registry, best practices).
- ChanersGenAI EngineerTECHJanuary 2024 - October 2025 (1 year and 9 months)Orsay, FranceDesign and deployment of generative AI and autonomous agent solutions, covering the entire back-end → front-end cycle: FastAPI/Docker microservices, industrial RAG pipelines, automated CI/CD, and production observability.Flagship projects:QuickResolve– Microservices RAG Platform (Architect & Full-Stack Lead)
- Complete ingestion and search chain: multi-format parsing (Docling), PII anonymization, chunking, Gemini embeddings, Qdrant vector storage, contextual chat.
- Stack: FastAPI (microservices), MinIO (S3), PostgreSQL, Next.js/React front-end, GitHub Actions CI/CD.
- Modular services: ingestion, orchestration, PII redaction, embeddings, conversational agent, parsing, data-generator.
AI Newsletter Generator– AI Newsletter Generator (AI Engineer & Full-Stack Lead)- Automated writing with OpenAI API / LLaMA coupled with real-time search (SERP / Perplexity).
- Hybrid RAG pipeline (Solr BM25 + dense embeddings + Qdrant).
- Stack: OpenAI API, LangChain, FastAPI, Next.js, Python.
TheMeetingReport– Intelligent Meeting Transcriber (AI Engineer & Full-Stack Lead)- Multi-speaker transcription & diarization with Whisper.
- Summaries and action generation via LangChain + LLMs.
- Advanced observability: end-to-end tracing with OpenTelemetry.
- Stack: Whisper, LangChain, FastAPI, PostgreSQL, Python.
- GeoFlexSoftware EngineerTECHJanuary 2019 - December 2023 (4 years and 11 months)Massy, FranceManaged the development of a real-time GNSS platform and critical software solutions for the automotive and telecommunications industries. Responsible for microservices architecture, CI/CD pipeline, and cloud & edge deployment (Docker/Kubernetes).Flagship projectsReal-time GNSS Platform(Lead Dev & Architect)
- Design and industrialization of a high-precision GNSS correction platform.
- Python/C++ microservices, FastAPI APIs, Docker/Kubernetes containerization, Prometheus → Grafana observability.
- Robust CI/CD pipeline under GitLab CI: automated tests, integration, deployments.
Renault Project– Sensor Fusion & GNSS Post-processing (Backend Lead Dev)- Development of containerized Python/Flask & C++ modules for GNSS processing + real-time sensor fusion (ROS node).
- Post-processing automation, mentoring a front-end intern.
- Results: scalable platform integrated into the Renault chain, successful knowledge transfer.
Orange Project– Real-time GNSS Calculation on Cloud (Lead Dev & K8s Integrator)- Migration of GNSS calculation to the Orange Business cloud.
- Kubernetes integration, monitoring scripts (latency/costs), technical presentations to BUs.
- Results: validated architecture, controlled latency, cross-team adoption.
Impact- Reduction of GNSS processing latency.
- Deployment of scalable, observable solutions compliant with industry standards.
- Technical leadership: code reviews, mentoring, dissemination of DevOps and ROS best practices.
Recommendations
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
- Master 2Aix-Marseille University2016Commmerce en Instrumentation Scientifique
- Master 1Paul Sabatier University2015Electronique des systémes embarqués et télécommunications