Yonk Shi

Background

About Yonk Shi

I’m a machine learning engineer with a background in academic research and a strong foundation in software engineering. I currently work at Kive.ai, a VC backed AI company building the AI quality engine behind Kive.ai’s hyperrealistic, studio-grade product photo + video generation.My academic work focused on representation learning for complex, real-world systems, and I’ve published at leading venues such as NeurIPS and ICLR. My research spans computer vision, speech, and language—areas where I remain hands-on and up to date. Whether I’m building segmentation models, prototyping audio classifiers, or evaluating multilingual transformer architectures, I aim to make models both effective and practical.Before my PhD, I worked as a software engineer, where I built robust, production-ready systems in Python, Java, and Node. That engineering background continues to inform how I structure ML pipelines, evaluate trade-offs, and bring research to deployment.I’ve worked in both startups and larger companies, and I’m comfortable navigating the different rhythms and priorities of each. I enjoy bridging the gap between business needs and engineering—turning open-ended ideas into clear, testable implementations. I also speak at meetups, guest-lecture at universities, and value clear communication as much as technical depth.Colleagues describe me as curious, dependable, and thoughtful. I was born in China, moved to North America as a kid, and have spent the past few years in Sweden—where I now call home. Outside of work, I’m big on sailing, outdoor adventures, and trying to convince my friends that cooking with MSG is just good science.(It is.)