Part XII — Building Real Products (End-to-End Projects)
Overview and links for this section of the guide.
On this page
The Projects
You've learned the theory, the prompt patterns, and the optimization tricks. Now we build real software. This part consists of three end-to-end projects that mimic common real-world AI applications.
- Project 3: Vibe Coder. An AI assistant that knows your specific codebase and helps you write features. This is "RAG for code."
- Project 4: Customer Support Triage. An automated routing system that classifies tickets, extracts data, and drafts replies. This is "classification + extraction."
- Project 5: Data Analyst. A tool that takes raw CSVs and answers questions with charts. This is "code generation + execution."
New Skills Unlocked
By building these, you will master:
- Repo-level context: How to feed a whole project into an LLM without blowing the context window.
- Reliable routing: How to trust an AI to send emails or assign tickets.
- Sandboxed execution: How to let an AI write and run Python code (Pandas) safely.
Where to go next
Explore next
Part XII — Building Real Products (End-to-End Projects) sub-sections
3 pages