I am a Postdoctoral Fellow with Harvard’s Programming Languages and Formal Methods groups and the Basis Research Institute.
I submitted my PhD thesis at the University of Melbourne, where I built neuro‑symbolic techniques that uses graph-based representation for neural networks for program analysis. Earlier, I spent 4 years at Cinnamon AI Lab working on visually rich document information extraction.
I build tools that treat code as both data and logic. My work lies at the intersection of software engineering, programming languages, and machine learning, coupling graph-based representation, large language models, and symbolic reasoning to make software development and formal proof more reliable and automated.
At Harvard, I am extending these ideas to proof automation in Lean and discovering drug repurposing candidates. At Basis, I am working on two projects: MARA and R-ADA