I am currently working on AI research agents for long-horizon autonomy over heterogeneous data.
Previously, I earned my PhD in Computer Science at UC Berkeley's RISELab, researching ML development automation.
Updates
Our paper Production Machine Learning Pipelines: Empirical Analysis and Optimization Opportunities has been accepted at SIGMOD ‘21!
Our vision for the future of dataframe research has been presented at VLDB ‘20 by Devin Petersohn. Paper: Towards Scalable Dataframe Systems. Video.
We presented our work “Demystifying a Dark Art: Understanding Real-World Machine Learning Model Development” virtually at HILDA 2020. Slides.
Our work titled “Demystifying a Dark Art: Understanding Real-World Machine Learning Model Development” has been accepted as a full paper and will be presented virtually at HILDA 2020.