Jafar B.

I build production AI systems for companies that need them to work reliably at scale: LLM agents and copilots, document intelligence, and the data pipelines behind them. Most of my clients hire me for months, not days. My current engagement, building AI copilots and agents for an enterprise client, has passed 2,900 hours. Across 31 Upwork jobs and 7,000+ hours, I've kept a 100% Job Success Score and Top Rated Plus status. WHAT I DO LLM agents, copilots and assistants: RAG, tool use, multi-step agent workflows, and integration with enterprise systems Document AI: OCR, structured data extraction and classification, including government-issued documents and Arabic text Computer vision: scalable inference pipelines (PyTorch, TorchServe) built for production traffic ML engineering and MLOps: deployment, monitoring and lifecycle maintenance on Kubernetes and cloud infrastructure Data engineering: data lakes, warehouses, and the pipelines that feed your models SELECTED WORK AI copilots and agents for an Upwork enterprise client (ongoing since 2025) AI-driven inventory system using OCR to extract data from complex sources (1,800 hours) ML-powered data extraction from government-issued documents for an enterprise client Distributed, cloud-based SaaS for document classification Scalable computer vision pipeline with TorchServe BACKGROUND PhD from the University of Twente, MSc in IT, and a BSc in Computer Science and Data Science (valedictorian, class of 2021). Based in Amsterdam, working in English and German. HOW I WORK Clients describe me as a clear communicator and a reliable collaborator. I keep you informed with regular progress updates, document what I build so your team can maintain it, and tell you early if something won't work. If you have an AI project that needs to get past the prototype stage, send me a message or book a consultation.