Hey, I'm Sidd!

I research machine learning systems for biomedicine. Currently, I'm building diffusion models for deconvolving tumor evolution, along with multi-modal health foundation models. I am advised by Dr. Russell Schwartz at Carnegie Mellon University, where we focus on evolutionary guided computational precision oncology.

My research interest centers around building effective AI copilots for biomedical researchers. On the model scale, how can we make efficient task-specific models that perform well, and are intrepretable (the tools)? On the systems scale, what does an effictive biomedical copilot look like (the toolbox)?