Harry Kabodha
Cofounder at Varosync; ML for drug discovery
hike
New York, New York
Joined April 2026
Summary
Entrepreneurial founder bridging machine learning and drug discovery: cofounder and CEO of Varosync, presenting work that applies physics-informed and geometric ML to practical problems in molecule design and safety. pmwcintl+1
Academic researcher in biomedical engineering who applies deep learning to protein dynamics and single-cell morphology during MS studies at Columbia University (Dumitrascu lab). columbia+2
Active participant and winner in competitive hackathon and pitch environments, demonstrating practical implementation and product focus (hackathon winner listings and student startup pitch participation). devpost+2
Focus on GPCR conformational dynamics and safety: develops tools (e.g., HYALINE) and methods to read 3D protein shapes and link structure to functional states to enable safer, more selective therapeutics. pmwcintl+1
Work
Education
Projects
Writing
HYALINE: Geometric Deep Learning for Accurate Prediction of G Protein-Coupled Receptor States
Work describing geometric deep learning approaches to predict GPCR conformational states to improve drug design and reduce off-target signalling.
Committor Learning with Strategic Sampling Unlocks Transient Binding Sites in GPCRs
Poster/abstract-level contribution addressing computational strategies for characterizing transient binding sites in GPCRs using learned committors and sampling techniques.