At LEET Solutions, I built PileTracker, a 3D computer vision system that estimates fertilizer stockpile volume from multi-view imagery using photogrammetric reconstruction and Open3D point-cloud processing. I also developed real-time YOLO perception for industrial video streams to detect equipment defects, PPE violations, and restricted-zone intrusions.
I've also built a self-hosted HR chatbot with RAG, deployed GPU inference with vLLM, and delivered backend microservices and REST APIs. Through 42 Lyon and 1337, I've completed 20+ hands-on projects across C, C++, Python, and JavaScript, and ranked 29th globally in the CodinGame Fall Challenge 2023.
