Photon (https://databricks.com/product/photon)

One of the TLs in Dremel/BigQuery query engine. Led and launched many features & improvements including:- Planner and runtime optimization to improve query performance- Distributed algorithms to make query scalable- Data manipulation language

Work in Chrome backend team, support not only Chrome, but also 100+ other Google applications (Search, GMail, Youtube etc.) to collect, process and analyze stability data uploaded by all Google users from all platforms & devices. We started to support third-party developers after Firebase was announced in Google I/0 2016.My contributions were mainly in system scalability and Dremel query latency optimization,- Invented an algorithm and designed a MapReduce pipeline to optimize the data organization, which made the Dremel query 3x+ faster and saved 5x+ resources in the meanwhile.- Designed and implemented the streaming pipeline to reduce data ingestion latency for Firebase (from data uploaded until searchable) from ~20 minutes to ~2 minutes.

Work in Bing Image Search Indexgen team,- Built the first superfresh pipeline for image search, reduce fresh image index latency from days to minutes- Owner of the image crawl & processing pipeline. Optimized the whole system to deal with billions of images every day, which is 10+ times more than before. Built distributed & large scale services to provide comprehensive image understanding features to other teams.- Release manager for a small team with 10+ devs, taking care of code integration, deployment & monitoring to make sure the release quality. Also, admin of the code branch to manage the build environment, tools & testing work, which improve whole team's engineer excellence & efficiency.- The key person to troubleshoot and handle most of the live site issues of image indexgen.

Work in Microsoft Research Asia,- Architect of online infrastructure of web-scale object level search engines, like EntityCube (http://entitycube.research.microsoft.com/) and Microsoft Academic Search (http://academic.research.microsoft.com/)