At Cloud Zeta, I drive the vision and development of the cloud based 3D data platform, focusing on OpenUSD for the web. We aim to transform how professionals manage and collaborate with 3D data, make it more efficient.

I lead the immersive streaming org at Magic Leap, which covers multiple products include: Remote Rendering, Omniverse, Multimedia Framework and Camera Framework.

Fun exercise together!

I joined the Google Daydream in early 2016 to work on immersive and ambient computing. I believe immersive computing is the next generation computing platform which will be more disruptive than mobile computing. My job at Google has evolved multiple times as the product matures and the marketing shifts. Overall, it involves the following areas: 1) leveraging on-device machine learning; 2) revolutionizing computer-human-interaction; 3) improving generic operating system.More specifically, my roles are/were (in a rough reverse chronological order):- TL/M for Google Assistant on Glasses- TL/M for Daydream SysUI and immersive experiences.- TL/M for immersive experiences on Google Glass Enterprise.- TL for Daydream Platform for Enterprise XR.- TL for DaydreamOS (the Android-based OS variant for HDMs).- TL for WorldSense 6DOF Positional Tracking on VR HMDs.- TL for Daydream VR video and multimedia stack.- Key contributor for Android Graphics and Camera stack for ARCore and Daydream VR.- Key contributor for Daydream VR SDK, Graphics, and Compositor.I may have also worked on a couple of secret projects, of which I can neither confirm nor deny the details without explicit approval from Google LLC

Fun fact: I was one of the six engineers who launched Vessel (with a complete backend, consumer-facing web frontend, creator-facing CMS, iOS, and Android apps). That was an awesome team.This was a great adventure in the true startup land. At the frontline of building a startup from ground up, we went through prototyping, fundraising, launching, pivoting, etc.As the video, content, and encoding guy in the company directly reporting to the CTO, I was responsible for the following components in our tech stack:- Creator portal (content manager system): As the leading full stack software engineer, I'm in charge of designing, architecting, implementing, and implementing both the content backend (AWS, MySQL, Ningx, CherryPy) and the creator-facing React.js based website.- Video Ingesting: 1) serving content creators from all over the world to efficiently uploading high quality video content (a good chunk of those creators had shitty ISP); 2) helping creators to automatically migrating content from Youtube; 3) automated content ingestion pipeline from major content partners (e.g. Vevo, BuzzFeed, etc).- Video Transcoding: I built the in-house video encoding pipeline based on AWS Lamba (a server-less pipeline), which saved the transcoding cost by 97% (less than $0.10 per hour of video encoding) compared to the previous Zencoder-based solution.- Video Playback: implemented Vessel's in house pure HTML5 video players for all of our web products (main site, creator portal, mobile site).- Winner of "Techie Award" during the first Vessel Hack-days.

I was a core member of Hulu's content platform team. During my tenure at Hulu, I went through significantly growth in myself & my role (from tech newbie to capable TL leading major and complicated projects), and my team (from 3 engineers to ~20 engineers).- As project lead, I rebuilt Hulu's Distributed File System from ground up to meet the rapid growth in Hulu's content platform.- As project lead, I designed and implemented Hulu's partner-facing content ingestion system capable of handling 100TB+ data per day. The system was designed to be linear scalable and handles Hulu's incoming content years after I left the company.- As project lead, I built the first Hulu Live Streaming system in house, which became the foundation of Hulu TV years later.- Other contributions including video transcoding pipeline, CDN backend, data center monitoring system.That was the pre-cloud days. From today's perspective, that was almost in the dark age. I learned quite a lot about about to work with a real data center during that experience.

I was one of the lecturers (or lab lead) for UCLA CS35L (Software Construction Lab, https://web.cs.ucla.edu/classes/winter12/cs35L/). We were paid as an TA, but our actual job includes everything a proper lecturer fulfills:- Lecture planning, preparation and research.- Conducting laboratory sessions and giving out lecturing two hours a week (for an entire year).- Answering students questions in the lab and in the TA session.- Grading homework.- Designing and grading exams (each quarter needs to have completely different set of questions for the final exam).- Determine the final grade for each student based on assignments and final exam.- Guided by Paul Eggert, who is a legend in GNU/Emacs community.