Rohil Badkundri

I'm interested in developing AI systems to accelerate scientific progress.

I'm currently working on using AI to forecast clinical trial outcomes. I believe this to be a critical bottleneck in curing disease as we build more powerful AI systems.

Previously, I was on the early team at EvolutionaryScale, where we studied the scaling hypothesis for biological foundation models.

Some of my previous work:

• Co-led protein structure prediction model training at EvolutionaryScale
• Helped develop ESM3, a 98B-parameter multimodal protein language model featured on the cover of Science
• Developed protein language models to design better gene therapy delivery at Dyno Therapeutics

I studied Applied Math and Economics at Harvard College.