Tan Zhang

Started working on cloud compute platform.

Led the core infra team at Moveworks: - Built a large-scale customer data infrastructure to process millions of documents for retrieval-augmented generation (RAG) to power enterprise search products. - Developed dynamic configuration infra to customize run-time functionalities of all the micro-services at Moveworks. This is a fundamental platform to support agentic workflow automation. - Deployed and maintained various core infrastructure components (e.g., mesh networking, Airflow, Kafka) and run-time server framework and libraries.

- Lead the T&S moderation platform team of 10 engineers: building large-scale moderation platform to detect and take down harmful contents from Snap.

Tech lead for the virtual currency platform of Snap Tokens and Snap Crystals, which empower in-game purchases, content creator tipping, etc. https://support.snapchat.com/en-US/a/snap-tokens

1. Ads serving/quality infrastructure for Google Local Services. Press: https://searchengineland.com/google-local-services-ads-30-cities-285895 2. Cloud based packet core infrastructure for LTE and Google WiFi Press: https://www.rcrwireless.com/20170314/telecom-software/google-expands-telecom-sdn-support-partners-with-bharti-airtel-and-sk-telecom-tag2 https://www.sdxcentral.com/articles/news/google-uses-vepc-deliver-wifi-train-stations-india/2016/04/

Wisconsin Wireless and NetworkinG Systems Lab (WiNGS): Worked on networked systems over unused TV spectrum (aka. TV whitespaces), spectrum measurement tools and infrastructure, vision analytic systems, mobile systems, and cross-layer optimization of networking stack. Details available at http://pages.cs.wisc.edu/~tzhang/.

Chrome Cloud Service Team: Worked on Flywheel - a mobile web compression proxy that has been serving millions of chrome users worldwide. Optimized the service backend to improve data savings.

Mobility and Networking Research Group: Worked on a wireless vision analytic system that leverages edge-computing nodes at camera sensors to selectively upload important video frames to the cloud for cross-camera image fusion and deeper object recognition.

Wireless Network Group: Designed and deployed a long-distance backhauling network over TV whitespace spectrum.

Wireless Network Group: Worked on TV whitespaces radio prototypes and designed a dual-band femto cell network spanning both the licensed cellular spectrum and unlicensed TV whitespaces.