I am a 3rd year PhD student at UCSD CSE, working with Hao Zhang at the Hao AI Lab. My research focuses on building efficient and scalable systems for LLMs — from serving and training to reasoning & agents.
Research Interests
- LLM Serving Systems — Disaggregated inference, goodput optimization (DistServe)
- LLM Training Systems — Disaggregated attention for long-context training (DistCA)
- LLM Reasoning & Agents — Efficient serving of reasoning programs (Dynasor/Certaindex)
- Database Systems — Data analytics platforms, query acceleration
Active Projects
| Project | Area | Venue |
|---|---|---|
| DistCA | Disaggregated LLM training | MLSys'26 |
| Dynasor/Certaindex | LLM reasoning serving | NeurIPS'25 |
| DistServe | Disaggregated LLM inference | OSDI'24 |
Experience
| When | Where | What |
|---|---|---|
| 2023 – present | UCSD | PhD student, CSE |
| Summer 2025 | Snowflake | Research intern |
| Summer 2024 | Microsoft Research | Intern, RiSE group |
| 2020 – 2023 | DataChat | Technical Lead — built core backend, NLP pipeline, led team of 5-10 engineers |
| 2017 – 2023 | UW-Madison | BS + MS in CS, advised by Jignesh Patel (Database Group) |
During Madison I also explored astronomy (Ka Ho Yuen), biochemistry (Senes Lab), and geoscience (Yuhao Kang) — 6 amazing years! 🎓
Talks
- DistServe: Disaggregating Prefill and Decoding for Goodput-optimized LLM Serving. CUDA Mode
Side Projects
- Alice in the Wondertech. blog