王
Wang Tong · 王通
LLM post-training and agent engineer, focused on model behavior, data feedback loops, and system boundaries.
I work on LLM post-training and agents, focusing on data generation, reward design, evaluation, and system troubleshooting.
These articles use equations, code, and logs to document training behavior, run artifacts, and the debugging process.
About this blog. The name comes from Mixture of Experts. I write about training data and algorithms, OpenVINO inference, remote GPU orchestration, Android debugging, and the tools used along the way.
Working on similar problems, or want to discuss an article? Find me on GitHub, LinkedIn, or by email.