Jerry Zhuo

I'm building a GxP Validation Copilot that uses LLMs, RAG, semantic search, and pgvector to automate traceability between pharmaceutical requirements and validation protocols.

At Catalyx, I design and automate validation workflows for FDA-regulated pharmaceutical manufacturing systems. I validate AI-powered machine vision platforms for OCR/OCV, barcode verification, artwork inspection, ROI detection, and defect classification.

Previously at Pennsylvania State University, I developed computer vision, robotics, and deep learning applications using NVIDIA Jetson Nano, ROS, TensorFlow, PyTorch, and OpenCV. I improved YOLOv5 inference from 6.5 FPS to 13.35 FPS and built autonomous navigation and industrial defect-detection systems.