Omschrijving

I'm a Ph.D. AI/ML engineer with 10+ years building and deploying enterprise AI at scale — for the United Nations, the European Space Agency, Microsoft, and Shell.

I don't just advise on AI. I've built it: LLM platforms serving 2M+ beneficiaries, GPU inference systems at <200ms latency, and ML pipelines processing 10TB+ daily. I know what works in production — and what doesn't.

Available for: AI strategy & audits, LLM deployment & optimization, MLOps setup, AI team coaching, and fractional AI leadership. I work in English, Arabic, Dutch, and French — across Europe, the Gulf region, and globally remote.

Branche-expertise

Talen

Nederlands
Tweetalig / moedertaal
Arabisch
Vloeiend
Engels
Tweetalig / moedertaal
Frans
Basiskennis

Voorkeuren voor werkplek

Kan op locatie werken

Den Haag (tot 50km)

United Nations
AI & Data Consultant
januari 2026 - Vandaag (5 maanden)
–Deployed and managed AI/ML systems at enterprise scale — building scalable inference pipelines and integrating vector databases for efficient data retrieval.
–Implemented monitoring and alerting systems for AI/ML workloads to ensure SLA/SLO compliance across 6 operational deployments serving 2M+ beneficiaries.
MLOps / LLMOps end-to-end Bridging deep tech with business strategy
Ali&Sons
AI lead
augustus 2025 - januari 2026 (5 maanden)
–Deployed and optimised LLMs on enterprise GPU infrastructure — building sovereign AI platforms using vLLM and Ollama for high-performance inference at <200ms p95 latency across 3 regulated business units.
–Integrated and managed vector databases (FAISS, Pinecone) and implemented CI/CD pipelines for model updates, reducing analyst cycles by ~60% across 5 production LLM systems covering 500K+ documents.
–Developed monitoring, alerting, and dashboarding systems ensuring SLA/SLO compliance across 12 enterprise AI use cases in 8 business units.
LLM optimisation & GPU infrastructure MLOps / LLMOps end-to-end
European Space Agency
AI engineer
april 2022 - augustus 2025 (3 jaren en 4 maanden)
–Deployed and managed LLMs and vision models on NVIDIA GPU platforms — achieving zero-failure reliability at ≤1MB memory in mission-critical environments.
–Optimised LLM inference performance achieving 4–8× improvements through benchmarking across hardware and software configurations while sustaining production reliability.
–Built and maintained scalable inference pipelines using Kubernetes and Docker, cutting integration timeline by 30% across 3 production AI systems.
LLM optimisation & GPU infrastructure Bridging deep tech with business strategy