01. Capstone Overview

This chapter establishes the roadmap for the entire course and clarifies what you will build. A capstone project succeeds or fails in the planning phase, long before a single line of code is written. The choices made here—product scope, target audience, technical constraints—determine everything that follows.

Starting With the End

Every product exists to solve a problem for someone. Before writing any code, you need to answer three questions: Who has the problem? What is the problem exactly? How would you know if it were solved? These sound simple, but most failed products skip directly to the third question and guess at the first two.

Consider a local AI writing assistant. The problem might be "people spend too much time rewriting content" or "non-native speakers struggle with professional tone" or "teams lack consistent documentation style." Each interpretation leads to a different product with different features. Only user research reveals which problem is worth solving.

Choosing Your Scope

Local AI products must make choices that cloud-based products do not face. You are responsible for installation, hardware requirements, updates, and troubleshooting. This creates friction that cloud products avoid but also creates opportunities for privacy, offline use, and customization.

A realistic scope for your first local AI product targets a specific use case with a well-defined audience. Avoid the temptation to build a platform. Build a tool that does one thing well for a specific person in a specific context.

Project Structure

Your capstone project will follow this sequence: plan the product, understand the users, design the architecture, select MVP features, implement the backend, implement the frontend, test everything, package for distribution, write documentation, deploy, and finally launch with feedback collection.

Each chapter builds on the previous one, so work through them in order. Skipping chapters because something seems unnecessary is a common mistake that leads to rework later.

Setting Up Your Workspace

Create a project directory and initialize version control before writing any code. Structure matters from the start.

mkdir my-first-ai-product
cd my-first-ai-product
git init
touch README.md
touch TODO.md
mkdir -p src tests docs

Use TODO.md to track your progress and decisions. This becomes a living document of your thinking as the project evolves.