I'm a senior software engineer at Google, living in the SF Bay Area. Currently, I'm building harness and memory systems for the Gemini agent. My work also covers its CLI tools, execution environment, and evaluation.
I became interested in AI as an undergraduate in 2017 and started building production AI systems at Google in 2023. As a founding engineer on Gemini in Google SecOps, I helped build a conversational AI assistant for security analysts. I later worked on an agentic context retriever for Google Workspace data, supporting Personal Intelligence in the Gemini app and Google Search.
I'm interested in how agents can complete long-horizon tasks reliably and learn from interaction, and in the scalable training and inference systems that support them. I want to connect execution and evaluation to model improvement, first for agents using CLI tools and computer interfaces, and eventually for models controlling physical actuators in robotics.
An experimental college project applying convolutional neural networks to occluded facial recognition, built while first studying artificial neural networks and backpropagation.
Internship and contracts with breaks, followed by full-time employment.
Illinois Institute of Technology · Chicago, IL
SGGS Institute of Engineering and Technology · Nanded, India
Ministry of Science and Technology competition, Government of India.