RESOURCES
Curated, not collected
When another source teaches something better than we could, we send you there. Every entry says why it earns your time, and official documentation is preferred wherever a framework is being taught. No link dumps.
Frameworks
- LangGraph documentation LangChain · docs
The canonical reference for the graph runtime Course 001 adopts in Module 03. Read persistence and interrupts before trusting either.
- LangGraph persistence LangChain · docs
Answers exactly what one checkpoint contains and which tables the Postgres saver writes. The primary source behind Module 03.
Agent runtimes
- Deep Agents documentation LangChain · docs
The harness Module 12 interrogates. Read the mechanisms, then locate each guarantee yourself before relying on it.
- Temporal AI cookbook Temporal · docs
What claim, lease, heartbeat and durable waits look like when a workflow engine owns them. The comparison target for Modules 05 and 07.
- Effective harnesses for long-running agents Anthropic · post
The industry statement of the harness/runtime split the whole curriculum is built around.
- How we built our multi-agent research system Anthropic · post
A production long-horizon system described with unusual candor. Extract the six primitives from it as an exercise.
Evaluation
- Creating a LLM-as-a-Judge that drives business results Hamel Husain · guide
The eval methodology Course 001 follows: error analysis first, binary criteria, judges validated against humans. The single most load-bearing external read in the curriculum.
- Frequently asked questions about evals Hamel Husain, Shreya Shankar · guide
Short answers to the objections your team will raise when you propose error analysis before metrics.
- Evaluating and improving Agent at scale Replit · post
Evals as the development loop argued by a team with millions of production runs.
Observability
- Langfuse documentation Langfuse · docs
Self-hostable tracing that satisfies the course's observability contract: exact prompts, spans, tokens, cost, run correlation.
Context/RAG
- Effective context engineering for AI agents Anthropic · post
The published basis for the claim that compaction alone is not sufficient and structured notes change agent behavior.
Coding agents
- Harness engineering OpenAI · post
The coding-agent harness architecture from the team shipping Codex, and a live example of repository-as-durable-state.
Distributed systems
- PostgreSQL documentation PostgreSQL · reference
FOR UPDATE SKIP LOCKED, transactions, and the primitives the whole durability story stands on. Read the locking chapter once, properly.
Deployment
- Render free instance limitations Render · reference
The exact constraints Module 14 designs around. Verify before relying on the course's numbers; hosting limits change without notice.
- Neon plans and limits Neon · reference
Connection and storage caps that shape the pooling and lease-scheduler choices in the reference deployment.