Summary
Byungsoo Jeon is a senior systems software engineer and Ph.D. graduate from Carnegie Mellon University with 11 years of experience building automated, portable distributed ML systems and compilers for efficient multimodal and large language model inference. He currently develops a compiler for distributed Transformer inference in TensorRT at NVIDIA, after helping build a distributed LLM inference engine at OctoAI. His research-driven approach focuses on parallelism, operator fusion, and graph optimizations, informed by internships at Google and AWS and mentorship from Prof. Tianqi Chen and Prof. Zhihao Jia. Comfortable bridging theory and production, he has a track record of shipping end-to-end systems from mobile games to tera-scale graph and tensor tools. Based in Pittsburgh, he blends deep ML systems research with hands-on engineering to make high-performance inference portable across hardware.
10 years of coding experience
10 years of employment as a software developer
Busan Science High School
Doctor of Philosophy - PhD, Computer Science, Doctor of Philosophy - PhD, Computer Science at Carnegie Mellon University
Bachelor of Science (B.S.), Computer Science, Summa Cum Laude, Bachelor of Science (B.S.), Computer Science, Summa Cum Laude at 한국과학기술원(KAIST)
Korean, English, Chinese