Projects for AI Engineers
AI Engineer projects are portfolio builds that apply the whole path to realistic problems: a RAG assistant with citations and evals, a tool-using agent with MCP servers, a document pipeline with structured output, an evaluation harness and a small fine-tuned model. They show that you can ship and measure, not only prototype.
Stage: Capstone and Career · Part of the AI Engineer roadmap · Last reviewed 19 November 2026
What to learn in Projects
- Ship a RAG assistant that cites sources and reports eval scores in its README
- Build an agent that calls tools through your own MCP server
- Extract structured fields from messy documents and validate them against a schema
- Create a regression suite that fails when answer quality drops
- Fine-tune a small model for one narrow task and compare it with a prompted baseline
Projects interview questions
How do you present a RAG project so a reviewer sees evidence of quality?
Include a test set of questions with expected sources, report retrieval recall and answer faithfulness before and after each change, show failure cases you found, and give cost and latency figures. Numbers and error analysis persuade more than a screenshot of a chat.
What is a good scope for a fine-tuned small-model project?
Pick one narrow, measurable task such as sorting support tickets into fixed categories or extracting fields into JSON. Build a labelled set, hold out a test split, compare against a prompted larger model, and report accuracy, latency and cost per thousand requests.
Where Projects fits in the AI Engineer roadmap
Projects belongs to the Capstone and Career stage (advanced level), which takes about 84 hours. Build what employers ask about.
Other units in this stage
Back to the full AI Engineer roadmap