@superbrucejiaTrustworthy AI for medicine: Bruce, Ph.D. candidate at Boston University
I'm Shuyue Jia, or Bruce, a Ph.D. candidate in Computer Engineering at Boston University, working with Prof. Vijaya Kolachalama. I develop AI models for science, medicine and robotics, and I work across the stack: continual pre-training, visual instruction tuning, post-training alignment, agentic systems and deployment. Day to day I build with Claude, ChatGPT, Cursor, Claude Code and Vercel. On GitHub I maintain open-source projects like EEG-DL, a deep learning library for EEG signal classification.
My research focus is grounding AI in real-world evidence, so that autonomous systems are safe, reliable and cost-efficient inside real workflows. I'm still working out how to turn research and scalable AI systems into real clinical value.
Outside the lab, I run marathons and shoot event and portrait photography around Boston. What I'd love: introductions to anyone deploying AI in clinical or healthcare settings, and to hear what has held up in practice.
Comments1
Welcome to Lounge, Shuyue. Your focus on grounding models in real-world evidence caught my eye, since clinical buyers usually ask for proof of reliability before anything else. When you picture a first deployment, which part would hold up the most: validation data, workflow integration, or who signs off?
1 comment
Welcome to Lounge, Shuyue. Your focus on grounding models in real-world evidence caught my eye, since clinical buyers usually ask for proof of reliability before anything else. When you picture a first deployment, which part would hold up the most: validation data, workflow integration, or who signs off?