Uses
Tools I work with
The software I reach for most often, and what I use each one for.
Data and BI
Reporting, dashboards and the models behind them
- Power BI
- DAX, Power Query, data models, row-level security
- SQL
- PostgreSQL, SQL Server, SQLite, DuckDB
- Excel
- Power Query, pivot models, VBA
- Tableau
- Client dashboards
Machine learning
Fraud, credit risk, churn and forecasting
- Python
- pandas, numpy, statsmodels
- scikit-learn
- Pipelines, calibration, evaluation
- XGBoost and LightGBM
- Gradient-boosted models
- SHAP
- Reasons behind each score
MLOps and delivery
Getting models into use and keeping them right
- FastAPI
- Scoring and data APIs
- MLflow
- Experiment tracking and the model registry
- Airflow
- Retraining and monitoring schedules
- Docker and Compose
- Repeatable environments
- GitHub Actions
- Tests and checks on every push
Apps and web
Interfaces people actually use
- TypeScript and React
- Analyst dashboards and apps
- NestJS and Electron
- Back ends and desktop apps
- Streamlit
- Quick internal tools
- Astro
- This website
Cloud
Where things run
- Azure
- Container Registry for model images
- Oracle Cloud
- Linux VMs for hosting