Human-centered AI division.

• Developed context-aware interaction techniques for AR multitasking and multiple app switching• Designed systems for AR language learning and AR shopping with recommender systems• Designed and trained neural networks for noisy EEG data• Led 6 experimental studies with 150+ participants, wrote 9 publications, 5 as lead author• Led project collaborations with UCSB Department of Psychology, School of Education, and Osaka University

• Developed VR data collection environment for correlating eye signals with cognitive states.• Conduct research correlating eye tracking signals with vocabulary word understanding.

• Built libraries for infrared eye camera calibration, tracking, and frame processing in AR/VR headsets.• Conduct research on optical eye tracking applications for AR and VR headsets.

Vuforia Computer Vision Group• Designed and trained CNNs for 6-DoF pose estimation of CAD models in RGB images• Extended company’s CAD model training pipeline with synthetic data generation for 3x faster training

Mobile Vision Team• Developed float quantization and graph pruning tools to optimize TensorFlow models (part of TF Lite)• Ported the Face Detection API in ML Kit (formerly Mobile Vision) to iOS• Benchmarked Face Detection model performance on iPhones

• Prototyping, R&D, and feature development for an AR remote assistance and video teleconferencing app (now known as Vuforia Chalk by PTC).• Implemented 2D tracking of hand-drawn annotations in video feed.• Implemented keyframe caching to accommodate network delays.• Created network performance measuring tools enabling team to optimize RTSP feed to hit target framerates.

Chrome Protector Team• Developed automated data analysis pipeline for detecting and classifying malware binaries uploaded to VirusTotal with machine learning• Built metrics dashboard used by whitehat hackers to identify malware characteristics in VirusTotal

• Researched and developed a custom gesture input system for touch screen fighting games trained via machine learning• Developed cognitive training games for iOS utilizing N-back to improve cognitive capacity and promote mental fitness

• Create learning materials and guides for online homework submission tools• Assist in grading and course management

Chromecast Team• Implemented predictive caching for Chromecast apps improving launch times 40% (released as “Fast Play” API)