I joined Microsoft Azure Research – Systems
(AzRS),
focusing on systems for AI efficiency.
- ➪ Reliable and efficient agentic workflow serving: Sherlock, Murakkab
- ➪ Fine-grained provisioning and scaling for generative model serving at scale: OpScale
- ➪ Efficient multimodal model serving: ModServe (input) and StreamWise (generation)
- ➪ Long-context LLM serving at multi-million token scale: Medha
I completed a PhD in Computer Science at
UIUC,
advised by
Prof. Ravishankar K. Iyer
.
My dissertation focuses on designing and integrating learning-based solutions seamlessly
(i.e., with
efficiency,
robustness, and
reliability)
into production cloud systems:
- ➪ [ATC24, AIOps24, SoCC24] Energy-efficient, SLO-aware interactive LLM serving
- ➪ [ATC23, NeurIPS23, MLSys24] Robust, at-scale ML model deployment in cloud systems
- ➪ [SoCC22, NeurIPS22] Multi-tenant serverless computing resource management
- ➪ [OSDI20, DSN24] Sustainable, intelligent, and resilient microservices resource management
I was a recipient of the inaugural 2023 ML Systems Rising Stars by MLCommons.
In the past, I have worked with IBM Research, Google SRG, Google DeepMind, and Microsoft Research. Before joining UIUC, I got my bachelor's degree in Computer Science from the University of Hong Kong. I worked with Prof. Heming Cui at the HKU Systems Software Group.
You can find my CV here (Aug. 2024).
Interests
- ML/LLM/Agent Systems
- ML for Systems / AIOps
- Serverless Computing
- Distributed Systems
Education
Ph.D. in Computer Science, 2024
University of Illinois, Urbana-Champaign
B.Eng in Computer Science, 2019
University of Hong Kong
Visiting Student, 2018
University of Wisconsin-Madison