Saarang Srinivasan

My name is Saarang Srinivasan and I’m a senior at Purdue University studying Computer Science. I’m fascinated by problems and have a strong interest in machine learning and algorithms. More specifically, I’m interested in designing efficient and scalable Systems for Machine Learning.

In the past, I’ve explored the TinyGrad ML framework, and developed a Translation Validation framework for it (still in the works). Formally, we take all generated GPU kernels of some compiled Tinygrad code, and we verify the equivalence any of the “rewrites” or “optimizations” that are further applied on this kernel.

More recently, my friends and I developed a GNN-based learned cost model to predict runtime of tensor programs. We later integrated it with the TVM’s existing AutoScheduler for tensor program optimization.

I’m an ex-Olympiad Gold Medalist at the Indian National Olympiad in Informatics. I most recently interned at Collins Aerospace where I worked on military flight software verification and testing using formal methods. Before that, I designed system architecture to ensure scalability and CI/CD for AI/IoT products using Kubernetes and Docker for Hacklab Solutions.

I’m looking for full time or internship roles in software engineering or machine learning from May 2025.