Google Ads (AI + Infra)

Building Google Ad Manager’s first AI Agent

Led a team of 5 in the end-to-end design and development of trythunderbird.com. Spearheaded the growth of Thunderbird's loan business, facilitating over $5M in loans through the platform to date.

Google Ads (Distributed Systems)

▪ Utilized distributed logical ordering to implement a C++ design that minimized expensive advertiser evaluations by 90%, reducing pipeline CPU usage by 60% and serving latency by 33%

▪ Engineered, designed, and fully deployed an automated system utilizing C++, RPCs, and Cron jobs to optimize a previously manual infrastructure workflow, resulting in a 50% apex latency decrease

▪ Empowered 150+ Ads engineers with critical insights by developing an internal dashboard that displays end-to-end Feed metadata latency utilizing SQL scripts

▪ Winner of the Ads Intern showcase event (1 of 3 Interns selected to present in front of Ads leadership)

YouTube

▪ Took ownership over a new live stream metadata viewer in Angular Typescript that is used across YouTube’s infrastructure teams to debug live stream playbacks

▪ Implemented a new C++ back-end in Google’s internal service framework that fetches all YouTube Live metadata points from a Spanner database using SQL

▪ Designed a full-stack integration between the front-end and back-end services via RPC

▪ Proactively fixed existing bugs and wrote unit tests to optimize system design and developer experience

▪ Collaborated on customer meetings while developing technical presentations to pitch Cisco services to clients (direct team was responsible for 25M in sales)

▪ Developed a module on Cisco SecureX that captures high-risk websites on a customer’s network using AMP endpoints and sends a potential warning message on WebEx Teams (Cisco Hackathon 1st place)

▪ Implemented a tool using Python and Kapicitor that leverages network automation to warn corporations about malicious security risks, decreasing malware detection latency by 12%