Capstone project
Capstone standard
What your capstone must include to pass, with example project directions.
Your capstone is a deployed, end-to-end solution to a realistic business problem. It brings together everything in the program: discovery, backend, frontend, AI, integration, deployment and maintenance.
What counts as a capstone
Your project must automate or support a complete business workflow with several connected parts, not a single AI call. A standalone chatbot or a basic "ask questions about a document" app does not meet the standard on its own. It needs real workflow integration, evaluation and deployment around it.
Example directions
Use these as starting points, or propose your own in week 9:
- Complaint handling. Classify complaints, extract key details, look up the relevant policy, recommend an action, get human approval, create a ticket and draft a response.
- Invoice processing. Extract invoice data, validate the supplier, catch duplicates, save to a database, route for approval and notify the right people.
- Recruitment assistant. Take in a job description and CVs, score candidates against it with evidence, build a shortlist and suggest interview questions.
- Customer service operations. Detect what the customer wants, find the answer in a knowledge base, resolve or escalate, update the system and report on trends.
What every capstone must include
- A GitHub repository with a clear README: the problem, how the solution works and how to run it.
- A business case. One page on the problem, who the users are and the value of solving it.
- An architecture diagram showing how the parts fit together.
- A working backend API and a frontend, deployed live with a public URL.
- Logging and evaluation evidence showing how well it works and how you'd know if it stopped working.
- Sample data so anyone can try it.
- A short demo video of the project in action.
Milestones
| Week | Milestone |
|---|---|
| 9 | Capstone proposal, built on your discovery and business case |
| 10 | Capstone running in production with logging and monitoring |
| 11 | Capstone checkpoint: end-to-end build with evaluation evidence |
| 12 | Demo day and technical validation |
Technical validation
After demo day, you'll have a one-to-one session where you explain, change and debug parts of your own code. This is how we confirm you understand what you built, whoever or whatever helped you write it. See the AI tools policy.
Finished this lesson?