Sreeja Kamishetty

Sreeja Kamishetty

Member of Technical Staff · Chrome AI Tech Lead · Chromium GSoC Founder · Angel Investor · GHC Speaker

About Me

Member of Technical Staff and Tech Lead at Google Chrome with 8+ years, currently leading the evaluation strategy for Chrome's Gemini-powered web agents. I architect eval infrastructure used across multiple agent teams to measure and improve LLM and agentic-trajectory quality. 2026 P&D Tech Impact Award recipient.

Grew up in Hyderabad in an entrepreneurial family that instilled the values I lead by: growth mindset, resilience, and service to community. Studied CS & ML at IIIT Hyderabad, published research, and filed a patent — all driven by a belief in tech's power to change lives.

Tennis and cricket since age 10 shaped how I think about teamwork and leadership. Vipassana meditation clarified my purpose: build technology that makes the world more equitable. As an angel investor, I back founders who share that conviction.

Google Chrome Mountain View, CA Angel Investor GSoC Founder GHC '23 Speaker Cricket & Tennis

My Values

Spiritual Optimism

Believing in the inherent goodness and potential of the human spirit to solve hard problems with compassion.

Honesty & Integrity

Upholding absolute truth and moral principles in architecture and team dynamics.

Compassion & Growth

Nurturing empathy and constant learning to bring out the best in everyone around me.

Work Experience

Tech Lead, Chrome AI Infrastructure — Gemini in Chrome

Aug 2024 – Present · Google Chrome · Mountain View, CA
  • Own the evaluation strategy for Chrome's Gemini-powered web-agent products, defining eval-correctness bar and rater-calibration standards now adopted across multiple agent teams.
  • Architected a hermetic state-management platform producing deterministic, reproducible eval runs — eliminated a top class of eval flakiness that previously blocked launch decisions.
  • Built a unified observability layer reconstructing distributed agent trajectories from Chrome and backend telemetry; cut mean-time-to-resolution for LLM debugging by 70%.

Member of Technical Staff — Chrome AI & Performance

May 2022 – Jul 2024 · Google Chrome · Seattle, WA
  • Led model-quality and evaluation workstreams for Chrome's first wave of LLM features — Auto Tabs, Compose, Wallpaper Search — from prototype to production launch.
  • Shipped 100+ ms of page-load latency wins on Chrome's critical execution path, contributing to measurable revenue impact across Search and Ads.
  • Designed the SpareTab architecture for faster cross-page navigation and built critical-path analysis infrastructure on top of Perfetto.

Software Engineer — Web on Android Performance

Jun 2019 – Apr 2022 · Google Chrome · London, UK
  • Shipped Multi-Page Architecture (MPArch) in Chromium — the foundational refactor that enabled Prerendering, BackForward Cache, Portals, and Fenced Frames.
  • Designed and launched BackForward Cache, accelerating back/forward navigation by 1,000 ms for an estimated 600M Chrome users.

Founder & Org Admin, Chromium Google Summer of Code

Jan 2020 – Present · Google / Chromium
  • Founded and ran the Chromium GSoC program for 5+ years; recruited and coordinated 200+ mentors, admins, and contributors worldwide.
  • Shipped 65 of 65 student projects to upstream Chromium — a 100% completion rate across program cohorts.

Publications

Patent · 2022

System and Method for Controlling Vehicular Pollution Concentration and Providing Maximum Traffic Flow Throughput

S. Kamishetty, S. Vadlamannati, P. Paruchuri · US Patent, 2022
Journal · 2020

Towards a Better Management of Urban Traffic Pollution Using a Pareto Max Flow Approach

S. Kamishetty, P. Paruchuri, S. Vadlamannati · Transportation Research Part D, 2020
Conference · 2020

Towards a Better Management of Emergency Evacuation Using Pareto Min Cost Max Flow Approach

S. Kamishetty, P. Paruchuri · VEHITS, 2020
Conference · 2018

A Prioritized Routing Assistant for Flow of Traffic

G. Gupta, S. Kamishetty, P. Paruchuri · IEEE ITSC, 2018
View full profile on Google Scholar

Blog

Weekly reflections on AI, engineering, and building things that matter.