Course Wiki
Build and Ship an AI-Assisted Full-Stack App — AI Dev Tools Zoomcamp Module 2
# Build and Ship an AI-Assisted Full-Stack App
Related Pages
AI Dev Tools Zoomcamp › Module 2: Build and Ship an AI-Assisted Full-Stack App › Build and Ship an AI-Assisted Full-Stack App
Notes
Build and Ship an AI-Assisted Full-Stack App
Overview
In this module, you build a working end-to-end application with AI assistance. The default project remains a small Snake Arena app because it is visual, interactive, and still requires the main parts of a real system:
- product spec
- frontend
- OpenAPI contract
- backend
- database
- tests
The goal is not to let an AI tool build everything unchecked. The goal is to practice a controlled workflow where AI helps you move faster and you verify each step.
The module ends with an app that runs on your machine: a frontend and a backend that talk to each other over a defined contract, with data persisted in SQLite. Everything needed to put it in front of other people — integration tests, containers, CI, deployment, and CI/CD — is Module 3.
You will:
- Write a small product spec with user stories, acceptance criteria, and non-goals before generating any code
- Draft a frontend prototype with an AI tool (Lovable, Bolt, Cursor, Claude Code, Codex, …), then pull it into a normal repo workflow and make it maintainable
- Define an OpenAPI contract as the source of truth between frontend and backend
- Implement a FastAPI backend against the contract, starting with a mock database and tests for the key endpoints
- Swap the mock store for SQLite, keeping the app database-agnostic so Postgres can replace it later without a rewrite
- Add unit and frontend tests that cover the behavior described in the spec and the contract
Reference app:
- https://github.com/alexeygrigorev/interview-canvas-share
Recording: Build and Ship an AI-Assisted Full-Stack App
Module summary: The companion article builds a collaborative system-design interview app from scratch. It moves from a frontend prototype with mocked calls to an OpenAPI contract, a FastAPI backend, WebSocket collaboration, and SQLite persistence, while testing each stage before moving on.
Read the article: Build and Ship a Full-Stack App with AI Coding Assistants
Module Deliverables
At the end of Module 2, your repo should include:
product-spec.md
AGENTS.md or equivalent
frontend/
backend/
openapi.yaml
tests/
docs/ai-usage-report.md
The app should run locally from the README, persist data in SQLite, and pass its own tests.
Homework
- Homework
Previous Cohort Materials
The previous version of this module is archived here:
Community Notes
Did you take notes? You can share them here.
- Add a link to your notes above this line
Key concepts
Related notes
- aidt-m01-ai-native-developer-workflow
- aidt-m03-test-containerize-and-deploy-an-ai-assisted-app
- aidt-m04-devops-and-observability-for-ai-built-apps
- aidt-m05-configuration
- aidt-m05-module-5-coding-agent-capabilities-mcp-skills-pl
- mlz-m01-crisp-dm