I connect business outcomes to model behavior and build the system in between.

Right now I'm a Senior Machine Learning Engineer at TRC Inc, working remotely from India. I build agentic AI systems, evaluation frameworks, document-intelligence pipelines, and 3D intelligence systems — work where plausible output isn't enough and the system has to be measurable, diagnosable, and reliable inside a real workflow.

My core interests are robotics, computer vision, and 3D perception: SfM at scale, Gaussian Splatting for real-time rendering, and LiDAR segmentation for utility infrastructure. That work taught me that reproducing a paper is only the beginning. Data quality, runtime, evaluation design, and deployment constraints often contain the harder research questions.

Today I apply the same mindset to LLM systems, and the loop runs in both directions: production failures decide which research questions are worth my time, and the experiments decide what ships next.