research-ml-template

Reusable template for research-oriented machine learning and data analysis projects.

This template is designed for:

• thesis projects
• exploratory data analysis
• research datasets
• lightweight ML experiments
• methodological validation
• reproducible notebooks
• reports and figures
• project health checks

Runtime

Preferred runtime:

• Python 3.11+
• uv
• pyproject.toml
• src/ layout
• pytest for tests
• ruff for linting

No Conda environment is created by default.

First setup for a new project

1. Copy or create a project from this template.
2. Rename src/project_name to the real package name.
3. Update this README.
4. Update docs/PROJECT_CONTEXT.md.
5. Add dataset information to docs/DATASET_CARD.md.
6. Initialize the uv environment.
7. Run the project doctor.

Important policy

Do not commit:

• real .env files
• large raw datasets
• trained model binaries
• notebook checkpoints
• local virtual environments
• cache folders

The template tracks structure, not heavy artifacts.

Validation

Run:

./scripts/doctor.ps1

For future project-level validation, use:

./scripts/validate.ps1