Objectives
- Implement the full app on your chosen architecture
- Persist data locally or in the cloud
- Integrate a Google API and a device hardware/software feature
- Test with unit tests and document your AI usage critically
Requirements
- Amended report: all Phase 1 deliverables, updated after feedback
- Form validation, event handling, exception handling and clear error messages
- One Google API: Maps, YouTube Data, Analytics, Cloud Storage, Drive, Calendar, Docs, Translate or Forms
- One device feature: camera, device info, location, sensors, microphone/speakers, a platform-specific feature, or inter-app communication
- Working persistent storage (file, local database or cloud)
- 3 unit tests with a results report
- AI engineering log. Without it, you lose marks for both the log and the implementation.
Tasks
- 01
Architecture in code
Apply the pattern from Phase 1. A monolithic project = zero for the whole phase.
- 02
Persistent storage
Local and/or cloud. Implemented but not running = half marks.
- 03
Google API
Pick one related to your idea. −2 if implemented but not running.
- 04
Device feature
−2 if implemented but not running.
- 05
Validation & errors
Validate every form and catch failures with a friendly message.
- 06
Unit tests
3 tests, designed, run and reported.
- 07
AI engineering log
For each significant AI use: prompt (verbatim) → output → your changes + how you verified them → why you used AI.
Deliverables
- Amended final report (one per team), with a contribution table on page 1
- Final application code on GitHub
- AI engineering log (appendix of the report)
Submission
- Upload the report and code to e-learning and hand in a printed report
- Include the signed coursework form and the Statement of Academic Honesty
Grading
| Criterion | Marks |
|---|---|
| Architecture implementation (group + individual) −2 if selected but not fully or correctly implemented. Monolithic = 0 for the phase. | 5 marks |
| Persistent data (individual) 5 implementation · 5 running. | 10 marks |
| Complete implementation + APIs (individual) 3 form validation · 3 exception handling · 2.5 Google API · 2.5 device API. | 11 marks |
| Unit tests (group) 3 per successful test. | 9 marks |
| AI engineering log (group + individual) 5 completeness · 3 prompt quality · 4 critical analysis · 3 rationale. | 15 marks |
| Total | 50 marks |
Important
Generic, copy-pasted prompts score low on prompt quality. Write your own prompts and include the module, constraints and specifics in each one.
Resources
FAQ
Can I ask AI to generate the whole app?
No. That is heavily penalised. Design the modules yourself, then use AI on specific, well-described parts.
What if AI code isn't in the log?
AI-generated code without its prompt in the log = 0 for the log and −5 from the Phase 2 implementation.
What if I can't explain AI code in the viva?
−2 from the AI discussion mark.