I’m Moe, a fourth-year PhD student in the database group at the University of Washington, Seattle. I work on applying of large language models to data discovery tasks such as table annotation.
Previously, I worked graph compression methods, which accelerate graph analysis 10–100⨉ while introducing a minimal approximation error. I have also worked on automated machine learning (autoML), which allows for selecting performant and robust machine learning models. As well as causal inference, which aids analysts in finding true cause-and-effect relationships in complex data.
My work has found real-life applications at places from large enterprises, such as Microsoft Research, to small startups such as Virta Labs.
|I’ll be on a panel on AI+DB at the NWDS 2024.
|CHORUS will be at the Table Representation Learning workshop at NeurIPS ‘23. Come say hi!
|Read our preprint on using large language-models (LLMs) for data discovery.
|Presenting Quasi-stable coloring at VLDB 2023: try the Julia package out.
|Presented at the AutoML workshop at KDD 2022.
|Interning at the Gray Systems Lab at Microsoft this summer, working on cardinality estimation.
|Selected as a Herbold Fellow for the year 2021.
|Working under Chi Wang at the Data Systems Group within Microsoft Research this summer.
|View a demonstration of causal inference on relational data with CaRL, which I presented at VLDB 2020.
|Read my letter in The Seattle Times regarding the administration’s (since retracted) plan to expel international students.
|Received the Outstanding Senior Award from the Allen School of Computer Science.
|Excited to be joining the database group at the University of Washington as a PhD student in September 2020!
|Causal Relational Learning will be presented at SIGMOD 2020.
|Selected as a Mary Gates Research Scholar
|Honorable mention in the national CRA Outstanding Undergraduate Researcher Award.
- via email: kayali @ cs dot washington dot edu.
- Twitter: @moe_kayali
- Old-fashioned mail to:
3800 E Stevens Way NE
Seattle, WA 98195
Google Scholar, DBLP, Semantic Scholar, ORCiD .
My Erdős number is 3.