Jun 2021 – Nov 2023 · 2 yrs 5 mos · San Francisco Bay Area

Jul 2018 – Jun 2021 · 2 yrs 11 mos

• Jack of all trades. Did whatever necessary to ship Display.land, an innovative service that let you capture your world with your mobile phone and turn it into a 3D model on the cloud. The service has since been discontinued. There are nice examples of display.land in action on YouTube, like this: https://youtu.be/QfVfF52b-G0
• My work spanned frontend and backend, and across all engineering teams. It touched mobile, web, and the cloud. Some examples:
• Implemented the first version of multiple-player networking serialization engine.
• Led the backend team and implemented the initial APIs for all social features of Display.land.
• Led a small team and revamped the entire CI/CD system. Reduced CI run time from tens of minutes to one minute.
• Worked with the computer vision team and improved their development productivity. For example, we drastically reduced the build time of the gigantic C++ stack by leveraging Bazel.
• Worked with the mobile team to improve their development workflow; built tooling around Microsoft AppCenter to make over-the-air delivery reliable, which shortened the mobile app release cycles.
• Led the overall mobile and web end-to-end testing effort. We used cypress, puppeteer, wdio, and detox.
• Worked with the web team to transition our react code base from Javascript to Typescript without interruption.
• Was the designated hitter for many thorny tech issues. Some examples: figured out why some customers kept loading the old versions of our webapp (bad CDN caching); debugged severe performance degradation of our computer vision code on machines with many cores (low-level lock contention in the linux kernel), etc.

Jan 2015 – Jul 2018 · 3 yrs 6 mos

• Member of a small team tasked by the CTO to solve the most impactful technical headaches throughout the company.
• Some examples:
• Rebuilt aging analytics data ingestion pipeline responsible for recording tens of billions of rows of data into multiple Vertica clusters each day. The old system was fragile and required significant man hours to maintain. The new system leverages managed services like Amazon Kinesis, S3, DynamoDB, SQS, and EC2 Container Service to improve reliability and minimize ongoing operational costs.
• Rebuilt global messaging platform responsible for sending hundreds of millions of emails and push notifications daily. The old system was too costly to update for the cloud and was preventing the company from discontinuing a punitive data center lease on time. Worked on a tiny team to identify and rewrite the critical pieces in Go and Java in record time by leveraging managed services like Amazon Simple Notification Service, Simple Workflow (SWF), ElasticCache, and Aurora, saving the company millions of dollars in unplanned capital expenditure.
• Identified and fixed critical security issues to enable a revenue-critical new title to ship worldwide on schedule. Hacked our own C++ game client to identify attack vectors. Built sandbox to test and plug known exploits. Analyzed large volume of game action data to identify new attack vectors. Built server guardrails to prevent runaway cheating.

Sep 2014 – Jan 2015 · 4 mos

• Led a cross-functional team in defining the company-wide machine learning infrastructure, which was an initiative mandated by the CEO.

Oct 2013 – Sep 2014 · 11 mos

• Owned the entire payments ecosystem that processed 95% of Zynga's revenue.
• Appointed by the CTO after the company-wide layoff to identify the biggest technical and staffing risks and to stabilize the team.
• Led a high risk and complicated operation to migrate the in-house PCI compliant payment processing data center to a cost effective third party solution, saving over $1M in operational costs year over year.

Feb 2013 – Oct 2013 · 8 mos

• Led a small team and guided the launch of CoasterVille, Solstice Arena, Battle Stone and Ninja Kingdom.
• Responsibilities included advising game studios on learned best practices, building shared tech to shorten dev cycles, and providing tactical support for game studios facing hard to debug issues or operational crises.

Jul 2011 – May 2013 · 1 yr 10 mos

• Member of a small team responsible for the technical health of a 20 million+ peak DAU and 1M+ peak concurrents game with 4000+ instances on AWS.
• Worked on everything from optimizing client and server performance, to improving ops readiness, to debugging complicated operational issues, to handling high pressure operational crises (for example, restoring the game in 12 hours after all compute and storage nodes were wiped out by a horrendous vendor bug).

Jan 2010 – Jul 2011 · 1 yr 6 mos

• Led the Systems team responsible for the long term health, scalability and performance of the largest social poker game in the world with 7M DAU and 500K peak concurrents.
• It was during the hyper-growth period of Zynga. It was chaotic and high pressure to have to respond to unrelenting threats from all directions.
• Some examples: fixing unannounced and late-minute Facebook platform breaking changes, plugging attacks that threatened to ruin our chip economy, and staying ahead of our growth curve that constantly threatened to bring everything down.