I'm an AI Engineer building production-ready AI systems, specializing in computer vision, VLMs, and MLOps. I love turning research ideas into practical products that solve real-world problems.

I like the messy part of AI, where a model works in a notebook and then has to survive real data.

• 🧬 Building vision models that are faster and more accurate for a specific problem, including custom YOLO architectures instead of off-the-shelf ones
• 🔬 Making vision and VLMs work on medical images, where mistakes matter
• 🚁 Teaching drones to see and react, from hand-gesture control to follow-me tracking in simulation (ROS 2 + PX4)

I enjoy fixing things in tools I use. I've had PRs merged into:

• ultralytics/ultralytics
• roboflow/supervision
• PaddlePaddle/PaddleOCR
• obss/SAHI

I also maintain pixpick, an interactive vision-input toolkit for CV frameworks (YOLO, SAM2, Supervision), published on PyPI with ~30K downloads.

Always happy to talk vision, VLMs, or open source: LinkedIn · Email