About bajucie
Practical expertise.
Practical expertise.
No shortcuts.
bajucie is a Wexford-based workshop platform built around one specific problem: ensemble machine learning methods in finance are taught theoretically almost everywhere, yet applied rarely. We do the opposite.
Founded 2025 · Wexford, Ireland
What this platform is actually for
Most courses on machine learning in finance stop at the model. They show you how to fit a random forest or tune an XGBoost classifier, then leave you with a notebook and no clear sense of what to do next. bajucie starts where that notebook ends.
Each workshop is built around a specific, bounded financial problem - credit risk scoring, factor selection for equity portfolios, regime detection in time series. Participants work through real datasets using scikit-learn, LightGBM, and SHAP, with structured assignments that force decision-making at every step.
The format is deliberately small. Cohorts stay under 18 participants so instructors can give meaningful feedback on actual code. Sessions run live, with asynchronous materials for review. There is no pre-recorded lecture library pretending to be a workshop.
What participants work with
- Random forests & bagging
- Gradient boosting (XGBoost, LGBM)
- Stacking & blending strategies
- Feature importance via SHAP
- Cross-validation for financial data
- Walk-forward backtesting logic
- Imbalanced class handling
- Model interpretability tools
Participants arrive with varied backgrounds - some from quantitative finance, some from data science, a few from risk management. The shared thread is that they already know the basics and want to close the gap between knowing and doing. If you are looking for an introduction to Python, this is not that. If you want to understand why a stacked model outperformed a single estimator on your specific dataset, this is exactly that.
The people behind the workshops
