Selected Work
Machine Learning Pipeline
A supervised and unsupervised modelling pipeline in Python, covering feature engineering, hyperparameter tuning, and cross-validated evaluation.
← All projectsHighlights
What Shipped
- Implemented supervised and unsupervised models over a shared feature-engineering stage.
- Tuned hyperparameters systematically rather than by hand, with cross-validation as the selection criterion.
- Structured the pipeline so preprocessing, training, and evaluation stayed separable and repeatable.
The Story
Behind the Build
Overview
A modelling pipeline in Python covering both supervised and unsupervised approaches, built over a single shared feature-engineering stage so the two families were compared on identical inputs rather than on separately prepared data.
Approach
Hyperparameters were tuned systematically rather than by hand, with cross-validation as the selection criterion — so a model was chosen on how it generalised, not on how it scored against one split.
Preprocessing, training and evaluation were kept separable and repeatable. Each stage can be changed or re-run without disturbing the others, which is what makes a comparison between model families meaningful rather than incidental.
Keep Exploring