Summary
Qi Sun is a researcher and software engineer based in Minato, Japan, with five years of experience focused on advancing large language model capabilities and production infra. At Sakana AI they lead “senior coach” work teaching LLMs new skills via evolutionary context search and collaboration, with multiple ICLR and other top-conference papers under their belt. Their background spans hands-on systems engineering—accelerating AlphaFold2 on Intel CPUs and building GPU clusters—to designing novel model-training techniques like evolutionary model merging and token-dropping. Currently pursuing a doctorate in computer science at Institute of Science Tokyo after an MEng from Tokyo Institute of Technology, they combine rigorous academic research with pragmatic infra delivery. Notably, they pair deep ML research with the willingness to “glue to a chair” debugging distributed systems until production works. They enjoy building interesting, useful tools that bridge cutting-edge models and real-world engineering.
5 years of coding experience
Master of Engineering - MEng, Computer Science, Master of Engineering - MEng, Computer Science at Tokyo Institute of Technology
Doctor's Degree, Computer Science, Doctor's Degree, Computer Science at Institute of Science Tokyo (Science Tokyo)
Bachelor of Engineering - BE, Computer Science, Bachelor of Engineering - BE, Computer Science at Dalian University of Technology