Part of the Generative AI unit working on personalization and user-memory systems for Gemini frontier models. Working on evaluation and post-training (SFT and RL) efforts.

Currently, part of the Generative AI unit at GDM working on personalization of LLMs.

Previously, was part of an applied research team with a focus on improving search and discovery across Google products using state-of-the-art research in ML, IR, and NLP. Was specifically working on personalized content generation, improving information quality, dynamic content understanding, and creator modeling.

Previously, built low-latency distributed and scalable data infrastructure to enable efficient and effective management of compute, storage, and network resources Google-wide.

Course: Computational Science and Engineering Algorithms

Responsibilities included assisting graduate students in understanding course material, resolving questions regarding assignments and course projects, and grading assignments, exams, and projects.

An Optimal Fuzzy Controller for Inverted Pendulum Problem

*Implemented a fuzzy controller for inverted pendulum problem and tuned it with different evolutionary algorithms.

*Studied time complexity for both the normal and the vectorized implementations of the algorithms and achieved around 7% time reduction on account of optimization due to vectorization.

Study of Elliptic Curves: Arithmetic, Geometric and Algebraic Perspectives