Shiyi Cao

Shiyi Cao曹诗怡

I am a third-year Ph.D. student at UC Berkeley EECS, advised by Ion Stoica and Joseph Gonzalez, affiliated with Sky Computing Lab and BAIR.

Previously, I was fortunate to be working with Zhihao Jia at CMU on accelerating distributed training. I obtained M.S. in Computer Science at ETH, working at SPCL with Torsten Hoefler. Prior to joining grad school, I had a great time at Shanghai Jiao Tong University where I earned a bachelor's degree in Computer Science.

I am mainly interested in automated optimization of ML computations on large-scale heterogeneous hardwares.

Toward self-improving AI systems

From human-engineered AI to AI-designed AI.

What does it take for AI to improve AI? Cheap compute to serve and train models and agents, capable models trained to perform complex real-world tasks, and autonomous agents that discover and learn from better solutions.

Hover a project to trace how it connects; click it for details.

AI Infrastructure

Efficient and scalable AI systems

S-LoRAMLSys '24 VTCOSDI '24 MoE-LightningASPLOS '25

Model Training

Training capable reasoning models and agents

Sky-T1EMNLP '25 SkyRLTech report '25

Autonomous Optimization

Agents autonomously discovering and learning from better solutions

S*EMNLP '25 K-SearcharXiv '26 Evolving Agent TeamsNeurIPS '26
Projects I led or co-led, as first or co-first author. Solid: completed work. Dashed: where I'm going next.

News

Oct '26 Excited to release Evolving Agent Teams (NeurIPS 2026)!
Feb '26 Excited to release K-Search for automated GPU kernel generation!
Oct '25 Excited to receive Amazon AI Fellowship!
Aug '25 Two papers (Sky-T1 and S*) are accepted as EMNLP '25 Findings!
Earlier highlights (5 more)Hide earlier highlights
May '25 Excited to announce the release of SkyRL, an RL framework for training Real-World Long-Horizon Agents!
Feb '25 Sky-T1 is featured in The New York Times, The Wall Street Journal, and The Information!
Oct '24 Two papers (MoE-Lightning and GraphPipe) are accepted at ASPLOS '25!
Oct '23 Released S-LoRA, a scalable system for serving thousands of LoRA adapters concurrently!
Aug '23 Graduated from ETH and joined UC Berkeley EECS!

Publications

Projects I led or co-led, grouped by research area.

Autonomous Optimization

Agents autonomously discovering and learning from better solutions

EAT score curves on AMD and NVIDIA kernels, Mac serving, and RoboDojo

Evolving Agent Teams

Shiyi Cao, Ziming Mao, Dacheng Li, Joseph E. Gonzalez, Ion Stoica

NeurIPS 2026

Agent teams · GPU kernels · Long-running optimization

Paper Code Gym
S* test-time scaling

S*: Test-Time Scaling for Code Generation

Dacheng Li*, Shiyi Cao*, Chengkun Cao, Xiuyu Li, Shangyin Tan, Kurt Keutzer, Jiarong Xing, Joseph E. Gonzalez, Ion Stoica

Findings of EMNLP 2025

Test-time scaling · Code generation · Hybrid selection · LiveCodeBench

Paper Blog Code

Model Training

Training capable reasoning models and agents

SkyRL logo

SkyRL-Agent: Efficient RL Training for Multi-turn LLM Agent

Shiyi Cao, Dacheng Li, Fangzhou Zhao, Shuo Yuan, Sumanth R. Hegde, Connor Chen, Charlie Ruan, Tyler Griggs, Shu Liu, Eric Tang, Richard Liaw, Philipp Moritz, Matei Zaharia, Joseph E. Gonzalez, Ion Stoica

Technical report · 2025

Multi-turn agents · Agentic RL · Efficient training

Arxiv Code
Sky-T1 teaser

Sky-T1: Train your own O1 preview model within $450

Dacheng Li*, Shiyi Cao*, Tyler Griggs*, Shu Liu*, Xiangxi Mo, Eric Tang, Sumanth Hegde, Kourosh Hakhamaneshi, Shishir G Patil, Matei Zaharia, Joseph E. Gonzalez, Ion Stoica

Findings of EMNLP 2025

Long chain-of-thought · SFT & LoRA · Sky-T1 · Open reasoning models

Paper Blog Code

AI Infrastructure

Efficient and scalable AI systems

Earlier work

Networking and distributed systems

Open-source Highlights

Open frameworks and models.

SGLang

Lianmin Zheng, Liangsheng Yin, Zhiqiang Xie, Chuyue (Livia) Sun, Jeff Huang, Cody Hao Yu, Shiyi Cao, Christos Kozyrakis, Ion Stoica, Joseph E. Gonzalez, Clark Barrett, Ying Sheng

NeurIPS 2024

Structured generation and fast serving for large language models.

Paper Code

SkyRL

Shiyi Cao*, Sumanth Hegde*, Dacheng Li*, Tyler Griggs*, Shu Liu*, Eric Tang*, Jiayi Pan, Xingyao Wang, Akshay Malik, Kourosh Hakhamaneshi, Richard Liaw, Philipp Moritz, Matei Zaharia, Joseph E. Gonzalez, Ion Stoica

AgentX Competition 2025 · 1st Place (Research Track)

SkyRL-Agent-v0: the first open-source models trained with multi-turn RL on long-horizon SWE-Bench tasks.

Blog Code

Talks

Invited talks & visits.

May '26

Invited talk at Recursive on self-improving agents for long-running optimization tasks (e.g., kernel generation).

May '26

Invited talk at the Hao AI Lab, UC San Diego, on K-Search.

Feb '26

Invited talk at the Jiaolab, UC Berkeley, hosted by Prof. Jiantao Jiao, on K-Search.

Feb '25

Invited talk at the SLICE Lab, UC Berkeley, hosted by Prof. Sophia Shao, on MoE-Lightning.

Feb '25

Invited talk at the SAMPL group, University of Washington, hosted by Zihao Ye, on MoE-Lightning and Sky-T1.

Blog

All posts →

No posts yet.