01
Coming soon
How LLM Inference Works
Tokens, context windows, the KV cache and batching: why a model call costs what it costs.
Track · 8 courses
AI engineer
Ship product features built on language models: how inference works, retrieval, agents and tools, evals, and running it all in production.
Who it’s for: Software engineers adding LLM features to a real product. No machine learning background needed.
The courses, in order
0 of 8 live- 1Vote
- 021Vote
Coming soon
Building with LLM APIs
Prompts, structured output, streaming, retries and token budgets, from one API call to a feature you can ship.
- 032Vote
Coming soon
RAG That Holds Up in Production
Embeddings, chunking, hybrid search and reranking, plus the freshness and access problems demos skip.
- 041Vote
Coming soon
Build an Agent from Scratch
Tool calling and the agent loop in plain code: tools, stop conditions, errors and memory.
- 054Vote
Coming soon
MCP from Zero
Give an agent your own tools and data with a Model Context Protocol server.
- 062Vote
Coming soon
Evals for LLM Apps
Golden sets, regression suites and LLM judges, so a prompt change never ships blind.
- 071Vote
Coming soon
Prompt Injection, Attack by Attack
How untrusted text takes over a model with tools, and the defenses that actually help.
- 081Vote
Coming soon
Design an LLM Gateway
Routing, fallbacks, caching, rate limits and cost tracking in front of every model call.
Every course stands alone and teaches what it needs. Take them in this order, or start with the one you need today.
Vote on what’s next