Kuan-Wei Huang

I am a member of the Self-Improving ML Group. We are creating self-improving AI systems that can improve their own performance over time.

I was a founding member at Pegasi AI, an early-stage startup. I led the applied research and development of our core ML systems as the corrective layer for alignment of Gen AI applications.

As a R&D team member, I participated in designing and architecting containerized workflows using Kubeflow Pipelines to orchestrate multiple ML models for enterprise-level loss forecasting and risk management.

Independently, I worked on projects in applying deep learning computer vision models to gravitational lensing images, where I led the development of codebase and published results as the lead-author in NeurIPS and ECCV.

I was a Research Assistant working with Professor Sergey Koposov. My research projects focused on applying ML, statistical modeling, numerical simulations, and high performance computing in studying the Milky Way structure and supermassive black hole early evolution.

I was also a Teaching Assistant for lab courses of experimental Physics for 3 years.

I was a Research Assistant working with Professor Tzihong Chiueh. My research projects focused on simulating quantum wave dark matter numerically in high performance computing environment.

This was where the story started.