Sam User

I've built production AI systems spanning RAG agents, multi-agent trading platforms, computer vision, and data engineering. My work combines Python, PySpark, LangChain, Hugging Face, PyTorch, FastAPI, and OpenCV.

For UAE Expat Navigator, I architected a RAG pipeline across 27 official UAE government documents, using ChromaDB, source deduplication, and PDF-level citations. I built a FastAPI service with Groq Llama 3.1 that delivered sub-second responses on AMD MI300X hardware.

I designed BullRun's four-agent options-trading workflow with 12 deterministic risk gates, hash-chained audit logs, and human-approved execution. I also led EcoVision's computer vision architecture, improving SegFormer vegetation segmentation accuracy to 81.0% and delivering YOLOv8 processing at 10+ FPS.

I'm completing a BSc in Computer Science focused on Big Data and AI at the University of Wollongong Dubai. I'm targeting AI/ML Engineering and Data Engineering work where I can build reliable RAG/LLM systems, real-time inference products, and production data pipelines.