Her research spans video-language models, diffusion-based AI, and prompt optimization for few-shot learning. At Stanford Vision Lab, she co-authored VIDEOSHIFT, a benchmark for distribution shifts in video AI, collaborating with Panasonic to analyze 1,000+ real-world videos. Her work at Korea University’s Machine Learning and Vision Lab led to a 6% improvement in few-shot classification performance and was presented at ICCV 2023.
Jihee has also developed robotic AI solutions, including an autonomous shirt-folding system using a Franka Panda robotic arm and a 3D object tracking pipeline for robotic manipulation at SIMTech Singapore.
At Kos, Jihee applies her expertise in AI-driven decision-making, medical imaging, and generative models, ensuring precision, efficiency, and real-world impact in next-generation wearable healthcare technology.