Hi, I'm Murphy Tian, a MLSys researcher. I received my bachelor degree from University of Toronto, Computer Science Department. I am fortunate to work with talented research advisors: Prof. JunCheng Yang at Harvard MadSys Lab on LLM serving systems and Prof. Jiaxuan You at U Lab on agentic reinforcement learning. I am now a senior research scientist in Tencent. I also spent a wonderful year as a research intern in Microsoft Research Asia.
Feel free to reach out to explore potential research opportunities.
Research
Focus areas
My research focuses on building more capable and accessible intelligence through two complementary directions:
Post-training and agentic systems
Exploring reinforcement learning methods for agent capabilities and frameworks for automatic agentic system generation and optimization.
Efficient systems for AI
Building inference optimization, scheduling, and resource management infrastructure to reduce the cost of serving large language models.
My long-term goal is to make AI tokens as inexpensive and dependable as basic utilities, so that advanced AI capabilities can become accessible to the broader public.
Latest updates
News
Our paper SWE-Bench Mobile has been accepted to the KDD 2026 ADS Track as a Main Conference presentation.
FreeInference is alive. Try it out at freeinference.org.
Our paper on open-source Asian-English sentence simplification (OasisSimp) has been accepted as an oral presentation at LREC 2026.