Make your project usable by someone else.
A documentation workshop for the Practical Data Scraping and AI App Development capstones. Use it when preparing your final handover.
Capstone handover workshop · Modules 10
Work through these modules
Module 10 · Deploy, observe, and iterateWorkshop resources
What you’ll practise
- Make local setup reproducible from a clean environment.
- Explain evaluation evidence, limitations, and dependencies.
- Document how to observe failures and recover safely.
What to submit
- A README and safe example configuration.
- An architecture sketch and evaluation summary.
- An operating guide with a reproducible example run.
Your project tasks
Write a reproducible setup
Document prerequisites, dependency installation, configuration names, and how to run the project locally. Provide a safe example configuration with placeholder values, and explain how to obtain credentials separately.
- Use synthetic or authorized sample inputs.
- Record the versions and commands needed to reproduce the project.
- Include an expected result so another learner can check setup.
Explain quality and limitations
For a collection pipeline, describe the record schema, validation rules, provenance, and rate limits. For an AI app, describe the retrieval source, evaluation set, refusal behavior, and known failure cases.
- Show the checks used to judge correctness.
- Document what happens when a source, model, or dependency is unavailable.
- Explain what the project cannot safely or reliably do.
Hand over and recover
Give the reader a short operational guide: how to schedule or deploy the project, observe failures, stop it, and recover safely. Remove credentials and unnecessary personal data before sharing.
- Submit a README, architecture sketch, and example run.
- Have someone follow the instructions from a clean environment.
- Record the issues they find and revise the handover.
Completion criteria
- A second person can follow the setup without receiving your secrets.
- The example run includes expected results and known limitations.
- The evaluation summary identifies the tests used and remaining failure cases.
- The guide explains how to stop, diagnose, and recover the project.