InternshipKey terms and common mistakes
The words you will hear in MLOps work and the mistakes that most often cost interns approvals.
Key terms
| Term | Meaning |
|---|---|
| MLOps | Machine Learning Operations |
| DVC | Data Version Control |
| Pipeline step | One stage in the workflow |
| Feature | Input variable for a model |
| Batch inference | Scoring data in bulk |
| CI | Continuous Integration |
| Data drift | Input data distribution changes |
| Docker | Container platform |
Common mistakes to avoid
- Not versioning data and models along with code.
- Deploying a model with no monitoring for drift.
- Training and serving with different preprocessing.
- No way to roll back to the previous model.
Want to go deeper? The free MLOps learning path has short modules, games and a certificate for each module.