Kun Yang

Hello! I’m Kun Yang, a Machine Learning Engineer at Google, where I build agentic AI systems for LLM based code optimization — from AlphaEvolve-based pipelines that automatically discover and validate code efficiency improvements, to coding-efficiency benchmarks used in Gemini post-training. Before joining Google, I was a Senior Data Science Engineer at Juniper Networks. I received my Ph.D. from the University of Virginia (UVA) in 2024, where I worked under the guidance of Professor Cong Shen. My research interests include agentic AI systems — especially building evals and RL environments for long-horizon, persistent tasks — as well as prompt optimization for large language models (LLMs), reinforcement learning, and federated learning. I welcome collaboration and discussions with those who share these research interests.

For a deeper dive into my research work and publications, check out the publications and cv pages. If you are curious about my life outside of work, take a look at the misc page.