Databricks runs massive scale cloud infrastructure, managing millions of VM's and thousands of production databases every day across more than 30 geo-replicated regions in multiple cloud providers. Running infrastructure at this scale with a small 200 person engineering team requires a constant focus on operational discipline. In this talk, I will review the key tenets of an operations mindset and discuss what processes and mechanisms were put in place to allow for reliable service engineering at scale. Topics include team/org structure, talent strategy, service architecture, and key engineering processes (postmortems, service health reviews, SLA definitions). This talk will be most relevant for teams or companies running SaaS services.
Patrick Wendell co-founded Databricks in 2013. Databricks builds software for large scale machine learning and data processing, and today has thousands of customers and more than 1000 employees. At Databricks, Patrick is a VP of Engineering, leading teams focused on machine learning and data science products. Before Databricks, Patrick studied distributed systems engineering as a graduate student at U.C. Berkeley. He holds a B.S.E in Computer Science from Princeton, class of 2011.