Delivering for Data Sciences
Delivering data science work means getting results into people's hands: dashboards and small apps built with tools like Streamlit, and basic model deployment as an API or a scheduled batch scoring job. This unit covers packaging, monitoring for drift and handing a model to engineers.
Stage: Data Products and Communication · Part of the Data Science roadmap · Last reviewed 19 November 2026
What to learn in Delivering
- Build an interactive dashboard with Streamlit or a BI tool
- Wrap a trained model in a FastAPI endpoint that returns predictions
- Save and load a model with its preprocessing steps together
- Choose between batch scoring and real-time serving for a use case
- Monitor input drift and prediction quality after deployment
Delivering interview questions
What is model drift?
Drift is the decline in a model's performance because the world changed. Data drift means input distributions shifted, while concept drift means the relationship between inputs and the target changed. Monitoring inputs and outcomes and retraining on a schedule are the usual responses.
When is batch scoring better than a real-time API?
Batch scoring is better when predictions are needed on a schedule, such as nightly churn scores, because it is simpler and cheaper to run. Real-time serving is needed when a decision must be made at the moment of a user action.
Where Delivering fits in the Data Science roadmap
Delivering belongs to the Data Products and Communication stage (advanced level), which takes about 5 hours. Impact over accuracy.
Other units in this stage
Back to the full Data Science roadmap