Kartik Desai

I’m an AI- and machine-learning-focused Bachelor of Technology student in Computer Science and Engineering, currently serving as a Teaching Assistant for Data Structures and Applications. I enjoy translating complex ideas into working implementations and clear guidance for others.

In personal research, I built “Hamiltonian Hyperparameter Dynamics (HHD),” a self-tuning neural network optimizer that co-evolves weights and hyperparameters using symplectic (Hamiltonian Monte Carlo) integration. I designed a unified optimization framework with a three-phase HHD-Unified method and benchmarked it against Bayesian Optimization and pure HMC baselines across 5+ benchmarks, validating theoretical guarantees and scaling with statistically-backed results across 11 datasets.

I also develop product-ready tools—from a voice-activated desktop assistant automation (“Athena”) using a PyQt interface to a full-stack “DesaiArts” e-commerce storefront using Next.js and Tailwind CSS. Alongside this, my algorithmic work includes comparative research on Binary Knapsack Problem optimization across classical, metaheuristic, and quantum approaches, and I actively build leadership through co-ordinating campus initiatives.