Feb 2026 – Present · 4 mos

• Building your autonomous builder swarm.

Apr 2025 – Feb 2026 · 10 mos

• ML Inference @ Google (Cloud 🫱🏼‍🫲🏾 Gemini 🟰 Vertex AI)
• Making LLM serving/inference for Gemini delightfully reliable & reliably delightful at a planet-scale.
• Oh, and economical too.

Feb 2024 – Mar 2025 · 1 yr 1 mo

• WAN @ Google
• Keep Google visible to the world.
• How? Manage, at scale, peering technologies & intermediate systems via SoTA observability & mitigation systems.
• Why? To enable you to watch cat videos while simultaneously allowing you to build a variety of workloads in the cloud (maybe ML inference, or anything else for that matter). This is the forcing function ensuring Google handles multitudes of traffic and workload types.

May 2023 – Feb 2024 · 9 mos

• Monitoring and alerting for GCP and Google-internal services.
• Successfully migrated all internal users from 2 disparate tools (15-year old (v1) and 5-year old (v2)) to the next-gen alert visualisation tool (v3) with zero downtime. As a result, I trail-blazed the adoption of the Google-wide experimentation framework in the Monitoring & Alerting space, showcasing its safety. All in the space of a quarter (for context: 4 years prior, a similar effort lasting 6 months resulted in Google's internal users using both the tools instead of the 15-year old tool being fully deprecated).
• Learned and wrote (almost 10k lines of) Typescript, for the first time in production.

Dec 2022 – May 2023 · 5 mos

• Docs / Editors @ Google Workspace
• Derisked observability concerns in a week via 30 code changes requiring approval from over 15 engineering leads and managers in partner dev teams. For context, this problem went unsolved for over 2 quarters before I'd joined the team. I fixed production notifications so that SRE isn't blind to customer pain.
• Biggest challenges overcome: Slow bureaucracy (by being clear and upfront to stakeholders about the changes being made), learning just enough of an obsolete language to solve the issue end to end.

Apr 2021 – Present · 5 yrs 2 mos

• Automated incident troubleshooting @ Google
• Built and facilitated Google-wide adoption of an automated incident troubleshooting tool; impacting multiple product areas by reducing mean time to mitigation from hours (or even days) to minutes.

Mar 2019 – Dec 2022 · 3 yrs 9 mos

• Serverless @ Google Cloud Platform
• Making Google Cloud's Serverless compute offerings (App Engine, Cloud Functions, Cloud Run) more reliable. Check out cloud.google.com/serverless (featured at Google I/O '22 - https://youtu.be/qBkyU1TJKDg?t=2556)
• Projects
• Automated stack turnups
• Enabled safe, slow rollouts of changes to over 500 databases. This changed the status quo of having customers affected globally for multiple hours by bad schema changes to not having such issues thereafter.
• Set up operational excellence processes that helped the team and dev partner teams to sustainably improve ops load. The SRE team was able to onboard 40% more partner teams whilst continually reducing overall operational load, and getting rid of un-actionable alerts.
• Built an automated incident root causing system (used Google-wide). This reduced mean time to response and mitigation from multiple hours to just minutes.
• Operations
• Experienced oncaller having handled numerous incidents, followed up with postmortems. I've led large scale incident responses spanning numerous days (most notably Log4J remote-code-execution vulnerability response).
• Improved tooling across products to speed up debugging and impact analysis, by automating toilsome tasks into scripts/notebooks, thus reducing efforts which took hours into a result that's achieved in seconds.
• Community
• Mentored 10+ FTEs/interns into their roles
• Translated numerous math lessons for over 1M+ primary school children in developing countries for Oppia (check out oppia.org)