Welcome to Nextbit Handbook!¶
This handbook describes how we work, which tools we use, and how we suggest using them to be fully integrated into our team. Before you start working on a project, please skim through all sections and set up all the relevant tools.
Important
If something is not clear, ask around: colleagues are here to help. Even better, open an issue to improve the clarity of the murky sentence or section.
If you want to contribute to this handbook, check Contributing.
- Values
- Your first day
- Project workflow
- Machine Learning Development Guidelines
- Before you start to build models
- How to reliably build models
- How to robustly evaluate models
- Do use an appropriate test set
- Do use a validation set
- Do evaluate models multiple times
- Do save some data to evaluate your final model instance
- Don’t use accuracy with imbalanced data sets
- How to compare models fairly
- Don’t assume a bigger number means a better model
- Do use statistical tests when comparing models
- Do correct for multiple comparisons
- Don’t always believe results from community benchmarks
- Do consider combinations of models
- How to report your results
- Merge requests workflow
- Git workflow
- Open source
- Remote working manifesto
- How to report a bug
- Working with big issues
- Permissions
- VPN
- Gitlab CI
- Git Large File Storage
- Testing
- Amazon Web Services (AWS)
- Manage multiple AWS profiles on CLI