A research journey shaped by systems, shifts, and questions of reliability.

I am a Senior Multimodal AI Researcher at Dolby Laboratories’ Advanced Technology Group in Bengaluru. My research spans multimodal and generative AI, mechanistic interpretability, content attribution and protection, robust model evaluation, domain generalization and adaptation, and efficient inference. Across these areas, I am interested in a common question: how can we build AI systems whose behaviour remains understandable, dependable, and practically useful outside controlled settings?

My research journey began during the final semester of my undergraduate studies, when I joined the Autonomous and Cyber-Physical Systems Lab at the University of Agder in Norway as a visiting researcher. Working across autonomous systems, edge AI, and sensor-driven learning introduced me to the gap between strong laboratory performance and reliable operation in unfamiliar environments. That experience shaped my doctoral research at IIT Hyderabad, where I studied the generalizability and adaptability of deep learning models under domain shift.

Alongside my academic work, research experiences at Microsoft Research India and Adobe Research exposed me to large language models, multimodal systems, generative modelling, and efficient inference from complementary perspectives. I now continue this journey at Dolby, working on research problems at the intersection of intelligent media, reliable evaluation, attribution, protection, and deployment.

My broader goal is to connect theoretical understanding with practical systems: models that can be examined rather than treated as opaque, evaluated beyond narrow benchmarks, traced and protected when they generate content, and made efficient enough to operate in real-world environments.