Harry Kabodha

Biography
Harry Kabodha cofounded Varosync to bring physics-informed machine learning to the hardest problems in drug discovery. His work focuses on conformational dynamics, functional selectivity, and isoform-aware molecular design domains where subtle structural differences determine therapeutic success or catastrophic failure. Prior to Varosync, Harry conducted computational research at Columbia's Irving Institute for Cancer Dynamics, applying deep learning to protein dynamics and single-cell morphological analysis.

---

Talk
AI Reading GPCR Shapes to Design Safer Drugs
GPCR drugs often fail because they cannot distinguish subtle onoff shapes of the same receptor. Hyaline is a geometric deep learning system that reads 3D protein structures to tell which states are active or inactive, helping teams design GPCR medicines with cleaner signalling, fewer sideeffects, and higher clinical confidence.

AI for Emerging Therapeutics Showcase:
Varosync
Varosync builds AI models that help drug teams design molecules and nanoparticle formulations that hit the right target without harming healthy tissue. By reading 3D protein shapes and linking them to real clinical and delivery data, we improve target selectivity, dosing, and safety from discovery through translation.