Aashutosh A V

I'm an MS Computer Science student at Georgia Tech, specializing in machine learning. My work sits between generative modeling, multimodal learning and ML systems, and it tends to share one theme: I care as much about how you measure a model as how you train it.

At C21U I work with Dr. Jeonghyun (Jonna) Lee and Dr. Rod Valente on autonomous LLM teaching agents for Socratic Minds, Georgia Tech's AI tutoring platform. Alongside the agents, I built the pipeline that codes student dialogue against the ICAP engagement framework, and the evaluation harness that checks it against human coders.

With Dr. James Hays I work on sparse-view novel view synthesis — building an evaluation framework for diffusion and feedforward models on ScanNet++, and probing temporal self-attention inside a view-synthesis diffusion model to find where its geometry breaks down.

Before Georgia Tech I was a Research Intern at Microsoft Research India, working on workload forecasting and placement optimization for Azure Cosmos DB, which led to a VLDB paper and a U.S. patent filing. I graduated from BITS Pilani with a CS degree and a minor in Data Science, where my thesis was on representation learning for vision-language models.

I'm currently looking for ML research and engineering roles. If you think there's a fit, or you just want to talk about any of this, please reach out.