Home › Courses › AI Engineering Capstone Projects

AI Engineering Capstone Projects

Seventeen complete systems to build and ship: coding agents, production RAG, voice, video, fine-tuning, observability, and safety

Each lesson is a whole project rather than a concept: the problem it solves, the architecture, the stack, and a worked implementation to read against your own. Six are agents that act on real software — a terminal-native coding agent, a research agent, a DevOps troubleshooter, a code-migration agent, a multi-agent software team, an issue-to-pull-request agent. Four are retrieval systems, from RAG over a codebase to an MCP server with a registry. The rest cover an end-to-end fine-tuning pipeline, an LLM observability dashboard, a speculative-decoding server, a realtime voice assistant, video understanding, a constitutional safety harness, and a personal AI tutor.

4 phases · 17 lessons · 17 worked examples · Early access

Take ML & AI Engineering first — this course builds on it.

Every phase, every lesson, every project

• Agents in Production (6 lessons) — free — End-to-end agent systems that act on real software: a terminal-native coding agent, an autonomous research agent, a DevOps troubleshooter, a code-migration agent, a multi-agent software team, and an issue-to-pull-request agent.
  • Capstone 01 — Terminal-Native Coding Agent
  • Capstone 05 — Autonomous Research Agent (AI-Scientist Class)
  • Capstone 06 — DevOps Troubleshooting Agent for Kubernetes
  • Capstone 09 — Code Migration Agent (Repo-Level Language / Runtime Upgrade)
  • Capstone 10 — Multi-Agent Software Engineering Team
  • Capstone 16 — GitHub Issue-to-PR Autonomous Agent
• Retrieval & Knowledge Systems (4 lessons) — Systems that answer from a corpus they did not train on: RAG over a codebase, multimodal document QA, a production RAG chatbot, and an MCP server with a registry.
  • Capstone 02 — RAG over Codebase (Cross-Repo Semantic Search)
  • Capstone 04 — Multimodal Document QA (Vision-First PDF, Tables, Charts)
  • Capstone 08 — Production RAG Chatbot for a Regulated Vertical
  • Capstone 13: Stateless MCP Server with Registry and Governance
• Models, Serving & Observability (3 lessons) — The model side of a product: an end-to-end fine-tuning pipeline, an LLM observability dashboard, and a speculative-decoding server.
  • Capstone 07 — End-to-End Fine-Tuning Pipeline (Data to SFT to DPO to Serve)
  • Capstone 11 — LLM Observability & Eval Dashboard
  • Capstone 14 — Speculative-Decoding Inference Server
• Voice, Video, Safety & Tutoring (4 lessons) — Four systems outside the text box: a realtime voice assistant, a video-understanding pipeline, a constitutional safety harness, and a personal AI tutor.
  • Capstone 03 — Real-Time Voice Assistant (ASR to LLM to TTS)
  • Capstone 12 — Video Understanding Pipeline (Scene, QA, Search)
  • Capstone 15 — Constitutional Safety Harness + Red-Team Range
  • Capstone 17 — Personal AI Tutor (Adaptive, Multimodal, with Memory)

The technologies you will use

Python · MCP · OpenTelemetry

Common questions

Are these projects graded?

Not yet. Each ships the architecture and a worked implementation to compare against, but nobody and nothing reviews your build. A project is judged against its own acceptance criteria, which needs a per-project review that has not been written. The course is marked early access for that reason.

Which project should I start with?

The first phase — agents that act on real software — is free, and the terminal-native coding agent is the most self-contained of them.

Key terms in this course

Agent · RAG (Retrieval-Augmented Generation) · Fine-Tuning · MCP (Model Context Protocol) · Speculative Decoding

Continue your learning path

AI Systems Build Tracks · Agentic AI Engineering · ML & AI Engineering

Start the AI Engineering Capstone Projects course

Create a free account — the opening phases of 29 of 35 courses are free, no credit card. Or see Pro pricing.

All courses · Pricing · About · FAQ · Glossary