I joined Google as a software director with the expectation of working on infrastructure projects, but I was given the opportunity to work some applications (consumer web proprties) as a way of better understanding their needs In my first 18 months I managed Google News, Froogle (shoppig), Video (when it was video search and before YouTube was created/acquired), Finance when it was developed and launched. I later (around 2010) worked image search when it launched its new UI. I also managed a browser plugin project that implemented the first versions of client side web acceleration such as preloading and URL prediction. These efforts to improve user latency grew into a larger project some years later including [Make the Web Faster] (https://developers.google.com/speed/), TCP improvements for faster connection startup time as well browser and JS serving changes.

In 2005 I transitioned to working on search infrastructure and was tasked with rebuilding and growing the team and systems. I spent the next 7 years growing this team from about 10 people to over 700 and on rebuilding all of the indexing and serving systems that ran google.com. The original manually updated system was replaced by a fully automated, continuously updating indexing and serving system with a real-time layer and a large third tier that effectively covered the entire web. Amongst many interesting and challenging projects, the two I enjoued the most were moving query serving to a flash-based system and moving indexing to the incremental, document-at-a-time system. Yahoo and Microsoft were worthy competitors and drove us to continually improve everything we were building.

The search infrastructure team always worked closely with the teams building the data centers, networks and software infrastructure since we were effectively their largest and most demanding customer. In 2012 I transitioned to working on core infrastructure and led those teams for just over a year.

In late 2013 I started an exploratory project to build a secure, easy-to-use RPC, naming system supporting a synchronised distributed storage system for mobile devices. This was an opensource project from the start and is available as vanadium and on github. This in an effort that I’m personally very proud of and I continue to support the opensource code base.

By 2016 I waa looking for a new challenge and ready to leave Google after almost 13 years and ended up leaving to join GRAIL.