I build reliable software at the intersection of engineering and machine learning, from clean APIs and distributed infrastructure to GPU scale optimization research.
Software engineer and deep learning researcher from Surat, India, studying computer science at Texas A&M.
I build reliable full stack systems and research bilevel optimization, implicit differentiation, and GPU scale experimentation.
My work has been published at NeurIPS 2025. I care about models that work and systems that hold up.
I study the model, data, interface, and infrastructure as one system. I want to know why it works, where it fails, and whether people can use it.
Computer science gives me leverage. Machine learning models complex behavior. Physics and statistics sharpen how I reason about uncertainty.
Engineering: full stack products, APIs, databases, testing, and observability.
Research: bilevel optimization, implicit differentiation, GPU training, and reproducible experiments.
Research makes models sound. Engineering makes them useful.
I focus on strong baselines, reproducible code, careful evaluation, and practical deployment.
A concise record of research, engineering, and the work currently shaping my trajectory.
Released the scalable gradient based contract design preprint covering bilevel optimization, implicit gradients, and GPU scale experimentation.
Coauthored gradient based bilevel optimization work for principal agent contract design, accepted to the GenAI in Finance Workshop.
Started research on bilevel optimization, implicit differentiation, HVP/CG solvers, and scalable self supervised learning experiments.
Built Java and AnyLogic simulation models for civil and environmental systems, including agent based and discrete event workflows.
Applied Bayesian networks, MCMC, inverse modeling, and Python/R analysis workflows to supply chain risk and uncertainty modeling.
Began the B.S. in Computer Science with minors in Physics and Statistics.