Architected a high-throughput, auto-scaling LLM platform with async task queues, optimizing GPU utilization across self-hosted models (Qwen, Flux, Seedance) and third-party APIs (GPT, Gemini)

Led the end-to-end development of automated manufacturing systems, robotic arm simulations in ROS, and scalable testing programs, significantly reducing costs and boosting production efficiency

Developed an automated, cloud-based smart detection system for machine monitoring