Build something real

Projects

5 projects to build once the lessons behind them are done. Each ends with something you can measure and show.

Intermediate project
Document AI pipeline
Compare ways of extracting data from PDFs and pick a pipeline using measured accuracy.
Builds on 1 lessonabout 35 min of reading first
- Document and multimodal AI
Intermediate project
RAG with an eval harness
Build a RAG system you can measure, and use the measurements to decide what to change.
Builds on 3 lessonsabout 1 h 30 min of reading first
- Document processing
- Retrieval
- Retrieval-augmented generation
- RAG evaluation
Advanced project
MCP-connected agent with state-based evals
Build an agent that uses a tool server over MCP and measure how reliable it is.
Builds on 3 lessonsabout 1 h 20 min of reading first
- Tool calling
- Model Context Protocol
- Multi-agent systems
- Agent evaluation
Production project
LoRA fine-tune vs prompt baseline
Fine-tune a small model on a narrow task and test whether it beats the best prompt.
Builds on 3 lessonsabout 1 h 10 min of reading first
- Model adaptation choices
- Fine-tuning
Production project
Serving and latency profile
Measure where latency and throughput come from when serving a small model.
Builds on 2 lessonsabout 1 h 5 min of reading first
- Cost and latency
- Inference and serving

Document AI pipeline