Projects
Augmented Language
AI-powered Smart Glasses
2021 - 2024
I manage a team of SWEs and prototypers building Translate and Transcribe for Augmented Reality glasses and beyond. Our work on enabling live translation in everyday glasses was the closing feature at the Google I/O 2022 keynote:
Let's see what happens when we take our advances in translation and transcription, and deliver them in your line-of-sight. — Sundar
- Modeling and Improving Text Stability in Live Captions CHI 2023
- Unmet needs and opportunities for mobile translation AI CHI 2020
- US2019/095430 Speech translation device and associated method 2017
Expeditions & VR View
Stereo 360 imagery on the web and beyond.
2016 - 2018
I created VR View (Google Developer Blog) to help webmasters embed interactive stereo 360 imagery. I also built the first versions of the Expeditions Creator tool, used by teachers to create expeditions for their classrooms (Google Blog).
WebXR (née WebVR)
Open-source design and development
2015 — 2017
When the web's VR ecosystem was still nascent, I helped to bootstrap it and make the WebVR API available even on low end devices like Google Cardboard. I built and maintained libraries like the WebVR Polyfill to provide head tracking and lens distortion in vanilla JavaScript. WebVR Boilerplate helped me build numerous creative VR experiences to inspire and help novices get started with VR on the web.
Google Tone
Transmit the URL of the current browser tab to computers within earshot, over audio.
2015 - 2019
When you press the blue megaphone button in your browser bar, the computer's speakers emit a combination of audible and inaudible sounds that any nearby machine with a microphone can pick up. The signal triggers a Chrome notification. Clicking it opens the transmitted URL in a new browser tab.
Covered on the Google Research blog, TechCrunch, Verge
- US9,882,658 Communicating Data with Audible Harmonies 2017
- US 2018/0063572 Methods Systems and Media for Synchronizing Media Content with Audio Timecodes 2016
- US9,811,311 Using Ultrasound to Improve IMU-based Gesture Detection 2014
Google Cardboard
An award-winning, inexpensive virtual reality viewer for smartphones.
2014 - 2017
I invented, prototyped, and built the magnetic input method for the Cardboard headset. The initial run was 10,000 units. There are now more than 20 million viewers out there.
- Magnetic Input for Mobile Virtual Reality ISWC 2015
- US11,269,022 Magnetic controller for device control 2019
- US D750074 Virtual Reality Headset (Cardboard VR Design Patent) 2015
Chrome DevRel
Mobile web & Web Audio developer advocacy
2011 - 2013
I was the lead DevRel Engineer during the Chrome for Android & iOS launches. I specialized in mobile web development, writing articles with millions of views featured on https://html5rocks.com. My secondary focus was on browser audio, including writing an O'Reilly book about the Web Audio API. Since then, I have been an avid user of the API, which came together in a number of open source projects.
- Telejam: From Low Latency to No Latency WAC 2022
- US2022/014570 Synchronized performances for remotely located performers 2021
Social Computing
Crowdsourcing and visualization projects
2010 - now
My Crowdforge research at CMU showed that Microtask markets like Mechanical Turk can be used to accomplish complex tasks. I then got a taste for the wisdom of the crowds, as well as their foolishness. Could random people on the internet be corralled to create a tech tree? Could they be asked to tinker with an evolutionary origami art system? Can people with opposing viewpoints be brought together to participate in friendlier debates?
- Crowdforge: Crowdsourcing Complex Work UIST 2011
- Usnea: An Authorship Tool for Interactive Fiction using Retrieval Based Semantic Parsing ACL 2020
- An Assessment of Intrinsic and Extrinsic Motivation on Task Performance in Crowdsourcing Markets ICWSM 2011
Physical Computing
Reality is underrated
2009 - now
I like building things with my hands. It's a nice remedy to staring at screens all day long. Here are some projects that require more than fingers typing on a keyboard.
- Crowdforge: Crowdsourcing Complex Work UIST 2011
- Usnea: An Authorship Tool for Interactive Fiction using Retrieval Based Semantic Parsing ACL 2020
Apple
iWeb and iWork.com
2007 - 2009
Developed software architecture for significant portions of iWork.com and implemented new features for iWeb and the iWork suite in Objective-C. Also:
- Created a text and object selection engine in JavaScript.
- Built a cross-browser graphics library using SVG, Canvas and VML.
- Prototyped new product ideas using JavaScript and Ruby on Rails.
- Wrote parts of the Microsoft Office document viewer for iPhone Mail.
- US2010/0095198 Shared comments for online document collaboration 2008