Thursday 3:30 p.m.–5 p.m. in Room 23

Using Python with the Google Admin SDK APIs

Chris Coykendall


Part 1 Tired of repeatedly pressing buttons or manually copying things into Google Sheets? This talk will demonstrate how to get started with bypassing the UIs for Google products and make repetitive work much easier. The focus will be for business use cases on G Suite primarily and will dive into practical automation using Python with the Google Admin SDK APIs, but the patterns and practices can be applied to most all Google APIs. Speaker: Chris Coykendall is a Site Reliability Engineer (SRE) at Google on G Suite infrastructure, and co-author of chapters in the upcoming O'Reilly book "The Site Reliability Workbook". He currently works with the Google Admin SDK and Google Device Policy teams to ensure they are highly reliable and scalable. Chris is an internal reviewer of Python code style at Google and has worked in various engineering roles at Google for 5 years. Part 2 A Deep Dive into Monte Carlo Tree Search, the algorithm behind AlphaGoZero Speaker: Monte Carlo Tree Search (MCTS) is an algorithm that searches over a game tree for the best move, balancing between exploring unlikely moves and reading more deeply into likely moves. AlphaGoZero used it to great success, using a combination of MCTS and deep learning to bootstrap from random moves to superhuman levels of play. In this talk, one of the lead developers of MiniGo (an AlphaGoZero replication) explains how and why MCTS works. The talk will also dive into details of a performant Python/NumPy implementation. Some familiarity with Python assumed. No prior knowledge of NumPy or machine learning is needed. Brian Lee is a software engineer at Verily, where he works on problems in the intersection of machine learning and healthcare. In his spare time, he shaves yaks. He is a Recurse Center alum and is mostly self-taught.