At KMUTT, I built a fraud detection model with Python and Scikit-learn, evaluating Decision Tree classifiers with Gini Index and Entropy. I also shared the findings with non-technical teammates.
I built a gold and multi-market algorithmic trading system that combines price-action math with FinBERT news sentiment. I validated it through backtesting, holdout testing, and live paper trading, then expanded it to a 15-symbol portfolio using correlation analysis.
For a family business, I built an electricity and water billing system for an 86-room rental property, replacing paper billing. I verified its calculations against real records and fixed production data issues at their root cause.

