Hyung-gun Chi | Google

Biography

I am a Research Scientist at Google Android XR — Multimodal AI for Glasses, building multimodal LLMs that run in real time on wearable hardware. My work bridges foundation-model research and the hardware that ships to users: from 3D action recognition (InfoGCN, 500+ citations) to speech foundation model compression (DiceHuBERT) to on-device LLMs for Ray-Ban Meta glasses and Ray-Ban Meta Display.

Previously, I was a Research Engineer at Meta Reality Labs and an AIML Resident (converted to ML Engineer) at Apple Siri Speech & Understanding. I received my Ph.D. in ECE from Purdue University, advised by Prof. Karthik Ramani. 17+ papers at CVPR / ECCV / TPAMI / Interspeech, 1,000+ citations, and 7+ patents in multimodal AI, on-device foundation models, and human activity understanding.

Research Interests

I work on bringing foundation models out of the datacenter and onto the devices people actually wear — glasses, phones, and AR/VR headsets. That means tight compute, memory, and power budgets, and a different set of research questions than cloud-scale LLMs: What does efficient multimodal pretraining look like when the target is a 4-bit model? How do you distill a large speech encoder without losing downstream transfer? What representations compose vision, audio, language, and sensor streams into a single on-device policy?

On-Device Foundation Models — multimodal LLMs, speech foundation models, quantization, knowledge distillation, KV-cache & operator-level optimization.
Multimodal Representation Learning — vision–language–audio–tactile representations, self-supervised learning.
Human Activity Understanding — action recognition, motion generation, trajectory prediction for AR/VR and robotics.

Education

Ph.D. in Electrical and Computer Engineering, Purdue University Aug 2018 – Dec 2023
Thesis: Towards Improved Representations on Human Activity Understanding. Advisor: Prof. Karthik Ramani.
M.S. in Electrical and Computer Engineering, Purdue University Aug 2018 – Dec 2022
B.S. in Mechanical Engineering, Yonsei University Mar 2011 – Feb 2017

Experience

Research Scientist, Google — Android XR (Multimodal AI for Glasses) Jun 2026 – Present
Research Engineer, Meta Reality Labs — Voice AI, Ray-Ban Meta glasses Sep 2025 – May 2026
AIML Resident / ML Engineer, Apple — Siri Speech & Understanding Jul 2024 – Sep 2025
ML Research Intern, Toyota Research Institute May 2023 – Aug 2023
CV Research Intern, Honda Research Institute Jan–May 2023 & May–Aug 2022
Graduate Research Assistant, Purdue University — Convergence Design Lab Aug 2018 – Dec 2023

Contact

📫 stnoah1@gmail.com — reach out if you’re working on on-device multimodal foundation models for wearables. Always happy to chat research or collaborations.