Project Chiang M-AI

Engineering my way to a 100-miler with Generative AI

A comprehensive R&D log documenting the use of LLMs, Python automation, and system engineering to prepare for the Hoka Chiang Mai 160km.
methodology
prompt-engineering
sports-science
system-design
project-chiang-m-ai
ai-coaching
Author

nakmuaycoder

Published

January 25, 2026

“Fine-Tuning the Fighter: Lumpinee Stadium vs UTMB”

Fine-Tuning the Fighter: From Nak Muay to Ultra-Trailer

Baseline Model: A physiology overfitted for 5 rounds of explosivity (Muay Thai).
New Objective Function: Maximize endurance over 160km while minimizing the “Injury” loss function.
The Optimizer: Generative AI.

Project Chiang M-AI is a live experiment in Transfer Learning.
I am using LLMs to re-weight my fast-twitch muscle fibers into a diesel engine. The goal? Solving a complex domain adaptation problem: taking a body trained on the “Lumpinee Stadium” dataset and deploying it to the “UTMB” distribution.

The Episodes