Pro-PT150 for ALS

Discovered 3 novel targets in ALS and designed a prodrug for the treatment of neuroinflammation in ALS patients; the work was later picked up by Palisades Therapeutics and is now backed with wet-lab evidence including RNA sequencing and clinical trial endpoints. Clinical trial evidence is derived from PT150 trials in non-ALS diseases.

LatentGlue

Developed a large-scale foundation model called LatentGlue for molecular glue representation learning. Benchmarked to be SOTA for Molecular Glue activity prediction (Spearman's Correlation and RMSE). LatentGlue can also be used for more generally given the nature of its representations output, including drug discovery.

ISEF '25 CBIO 1st Place

Of the ~7 million people who compete in science fairs annually worldwide, ~1800 are selected to compete in the International Science and Engineering Fair (ISEF). My project was the top project in the Computational Biology and Bioinformatics category in 2025 for my teams (3) work on Pro-PT150 for ALS treatment.

ISEF '26 CBIO 2nd Place

Of the ~7 million people who compete in science fairs annually worldwide, ~1800 are selected to compete in the International Science and Engineering Fair (ISEF). My project was one of the top projects in the Computational Biology and Bioinformatics category in 2026 for my partners and I's work on Molecular Glue discovery where we discovered 2 new TPD-based treatments for Parkinson's and Pancreatic Cancer using LatentGlue and then we delivered wet-lab in a simulated GI tract and verified non-cytotoxicity in two cell cultures.

USACO Gold Division

Completed rigrous algorithmic programming problems to rank in the top ~400 pre-college competitive programmers in the United States

Samsung Solve for Tomorrow Winner

Developed an AI model to predict multi-organ temporal metabolic responses in response to drug doses, winning $10,000 in Samsung equipment