After 4+ years with Scribd I have decided to move on and join Technical Operations at LivingSocial. I will still be working remotely from apartment in Toronto.
These past 4 years with Scribd were incredible for me: I’ve made many good friends here, I became better as a professional and as a person, and I tried really hard to give back as much as I could. I watched the company grow from a small startup to one of the top 100 web sites in the world. I will definitely be closely watching the company as it grows and reaches its new heights because Scribd became a really important part of my professional and personal life and I really care about its future. My work at Scribd is by far is the longest period I’ve ever worked for the same company (all my previous jobs were ~1-1.5 years long tops) and it is pretty interesting to reflect on what was done in the last years, how the company grew and I grew with it. Even though it is not a technical blog, here is a list of my most interesting projects at Scribd:
- Rails application deployment infrastructure – We’ve started this project with a few mongrel-based manually managed servers and grew our infrastructure to a large multi-cluster system tightly integrated with CI and monitoring. New infrastructure is fully-automated and is much more stable and performant.
- Caching Infrastructure – Built by me back in 2009 to help with quickly growing traffic needs the system (based on squid, nginx + custom modules) proven to be really performant, scalable and reliable and is still serving 90%+ of our traffic greatly reducing application cluster load.
- Multiple MySQL clusters with sharding – We grew from a single server to a large multi-cluster sharded database architecture and I was the person responsible for many parts of the process: from setting up the clusters to writing Rails code to support it (including writing a plugin DbCharmer to add multi-server support to Rails ORM).
- Large Hbase, Hadoop and Solr clusters for storage and indexing of a multi-terabyte texts collection – After we outgrew our simple sharded mysql cluster for texts storage, I’ve built a new Hbase-based solution that’s proven to be really reliable and scalable and is serving thousands of requests a second for the last few years.
So, thanks Scribd for all the opportunities they gave me and for the great journey we walked together. Good luck guys!
