Graduate AI researcher at the University of Hawaiʻi at Mānoa, in the Department of Information and Computer Sciences.
Prior experience in software product management.
Vision-language-action models (VLAs) for humanoid locomotion and manipulation at Prof. Igor Molybog Lab.
I am pleased to share that I received the NSF NRT-AI DESCARTES fellowship award.
Shared with Fortune magazine how our lab is utilizing the rent-to-own compute platform B3IQ.
I will join ʻIolani Artificial Intelligence Lab as a faculty member for their Honors Machine Learning class.
I will attend the ICML 2026 conference on July 6–11 in Seoul, South Korea. Please come attend our presentation in COEX Hall A on July 11, 2026 at 11 am.
Our next iteration of the paper Data-Efficient Curation for Multimodal Reasoning under Fixed Training Protocols will be presented at the Efficient Multimodal Question Answering workshop at ICML 2026.
I will present the Robotica workshop, a video-to-robot teleoperation pipeline for the Unitree G1 built with ʻIolani Artificial Intelligence Lab students, at the Kamehameha Schools Ed Tech Conference (CONNECT26) at the Blaisdell Center, Honolulu, on June 2–3.
I’m presenting preliminary results from ShepBERT at the University of Hawaiʻi Cancer Center × Google Cloud AI Research Workshop 2026. ShepBERT is a BERT-style clinical language model pre-trained on 230,000+ patient EHR records and designed to predict 1-year cancer risk, with a focus on Native Hawaiian and Pacific Islander populations, who face a disproportionately high cancer burden relative to the mainland US. View the poster (PDF).
A new paper, Data-Efficient Curation for Multimodal Reasoning under Fixed Training Protocols, describing our approach to balancing difficulty and diversity in sampling high-quality supervised learning datasets for multimodal models.
I will attend the NeurIPS 2025 conference on December 2–7 in San Diego, CA. Come join our presentation in the San Diego Convention Center, Upper Level Ballroom 6DE, at 11:15 AM – 2:00 PM PT on December 6, 2025.
Our team (Yosub Shin, Michael Buriek, Boris Sobolev, Vikas Kumar, Haoyang Xu, Samuel Watson and me) took first place at the DCVLR challenge at NeurIPS 2025.
Collaborative Research: MATH-DT Dynamical Models and Statistical Guarantees for Sparse Sensing in Nuclear Digital Twins, as part of the Prof. Andrei Klishin lab.
I joined the Department of Information and Computer Sciences at the University of Hawaiʻi at Mānoa as an M.Sc./PhD scholar.