Daniel Oosthuyzen
I build production-grade trading systems by fusing statistics, numerical finance, and machine-learning with robust software engineering.
Here you’ll get a glimpse of my domain knowledge in algorithmic trading, ML data science and my keen ability to apply innovative mathematical concepts in live data pipelines. Each file has been curated from my personal archive of past works to demonstrate examples of how I use various forms of data science techniques and expertise.
All project files referenced in this showcase are in formats: .py(python), .ipynb(jupyter notebook), .mq5(mql5: a c++ derivative)
All files can be found at:
https://github.com/thinkquant/quant-portfolio
📑 Table of Contents
<aside> 💡
This section is the mathematical spine. Here I demonstrate how unsupervised clustering, Bayesian volatility estimation, Fourier analysis, and LPPLS bubble modelling work together to transform raw market data into feature-rich, self-cleaning datasets. Techniques on display include K-Means → KNN regime filtering, FFT-linear hybrids, Monte-Carlo uncertainty bands, and Bayesian-ridge cones—skills essential for anyone building statistically grounded signals.
</aside>