I’m transitioning into Machine Learning / MLOps (Machine Learning Operations).
I build small, reproducible projects using Python, SQL, and simple automation.
- Flower Image Classifier (PyTorch) — Train and run predictions from the command line (top-K). Includes clear setup steps and an automated run check.
- Finding Donors (scikit-learn) — Supervised learning case study to predict >$50K income with model comparison and tuning.
Report (HTML): andigles.github.io/finding-donors-ml - Bikeshare Analysis (Python) — Command-line data analysis with a small sample dataset, tests, and CI (Continuous Integration).
- Dog Breed Classifier (PyTorch) — Command-line evaluation of pretrained models; prints per-model accuracy and writes
results.csv.
- Reproducibility: every repo runs from a clean clone
- Plain language: define acronyms the first time
- Evidence: tests, CI checks, and clear run steps
FastAPI • Docker • SQL (DuckDB, Postgres) • MLflow • Prefect
- LinkedIn: https://www.linkedin.com/in/andigles