Tianyu Li

About

I am an assistant professor in the UW-Madison Database Group. I obtained my PhD from MIT EECS working with Prof. Samuel Madden. Before MIT, I worked with the infamous Andy Pavlo at CMU, where I received my bachelor's and master's degrees in computer science.

I am recruiting 2-3 MS/PhD students for Fall 2026 in cloud, data processing, or AI systems. Please reach out if you are interested in chatting.

Research

New data processing paradigms like serverless analytics and agentic exploration challenge the performance assumptions and abstraction boundaries of the legacy systems today's infrastructure is built on. My research identifies the resulting inefficiencies and builds novel abstractions to bridge the gap. Current directions include:

• Resilient composition for serverless compute — the Composable Resilient Step (CReSt) abstraction and the DARQ runtime, which executes exactly-once steps using database techniques like group commit and optimistic concurrency control.
• Transactions for AI agents — redesigning the OLTP stack for long-running, speculation-heavy workloads that interleave millisecond database actions with seconds of model reasoning.
• The semantic data warehouse — persisting the expensive semantic work behind natural-language questions so that the workload becomes the schema.

Publications

Please see my Google Scholar Profile.

Teaching

• CS839: Next-Generation Data Systems, Fall 2026, UW-Madison
• 6.5830/6.5831: Database Systems, Spring 2026, MIT

Service

• CIDR 2027, Program Committee
• VLDB 2027, Program Committee
• SIGMOD 2027, Program Committee
• New England Systems Day 2026, Program Committee
• CIDR 2026, Program Committee
• VLDB 2025, Program Committee (Distinguished Reviewer Award)

Contact

Email: litianyu AT cs.wisc.edu

Office: MH 7580