References: Agents

Source material

Source curriculum (structural mirror, cited as further study):
• Full Stack Deep Learning, "LLM Bootcamp" (Spring 2023):
    Agents
  Guest instructor: Harrison Chase (LangChain), with Charles Frye,
    Sergey Karayev, and Josh Tobin
  Course page: https://fullstackdeeplearning.com/llm-bootcamp/
  Lecture videos: publicly available on the Full Stack Deep Learning
    YouTube channel
  License: bootcamp materials are published free to view but no explicit
    license (Creative Commons or otherwise) is published; lecture videos
    are on YouTube under standard terms.
  Required attribution: "Based on the structure of the Full Stack Deep
    Learning LLM Bootcamp (Spring 2023), by Charles Frye, Sergey Karayev,
    and Josh Tobin (fullstackdeeplearning.com/llm-bootcamp). This is an
    independent structural mirror in original prose; it reproduces no
    course materials, and Full Stack Deep Learning does not endorse it."
This lesson mirrors the structure of the corresponding bootcamp session (agents as a production
shipping topic). Clawdemy's lessons are original prose taught at a
strictly technical-primer level; agent autonomy, agent safety, agent
alignment debates, and similar contested topics are out of scope here.

Watch this next

• Full Stack Deep Learning, LLM Bootcamp: Agents with Harrison Chase (LangChain). The session this lesson mirrors. The recorded version brings the framework-author perspective with examples drawn from LangChain’s early agent work.

Going deeper

A short, durable list. Each link is a specific next step, not a generic pile.

• The ReAct paper (Yao et al., 2022). The original “Reason + Act” prompting pattern that pre-dates structured function calling. Worth reading once to understand where the agent vocabulary comes from; the pattern itself has been largely replaced by structured function calling in production, but the literature still uses ReAct’s terms.

• Anthropic’s tool-use documentation. Practical reference for structured function calling on the Claude API: how to declare tools, the response format, the loop. Equivalent docs exist for OpenAI and Google APIs; the conceptual model is the same across hosts.

• LangGraph documentation. The graph-based successor to LangChain’s earlier AgentExecutor. Worth reading as one well-developed example of how agent frameworks express the loop, the tool contracts, the iteration caps, and the observability hooks. The patterns here generalize across frameworks even if the specific library names change.

Adjacent topics

Where this connects inside the track and the wider curriculum.

• Augmented language models (lesson 4). This lesson is the deep dive on the tool-use side of L4. The four steps are the same; the only addition is “model decides when to stop.”

• LLMOps (lesson 7). The evaluation, logging, and observability discipline this lesson scales up. Agents need the same machinery, with trajectory-level expectations rather than per-call expectations.

• What’s next (lesson 8). The build-vs-buy spectrum named there, and the mix architecture, both apply to agent builds (most production agents use frontier hosted for the planner, smaller models for inner sub-tasks).

• Track 14 lesson 12 (Building agents: the loop, observation, tool calling). The using-side companion: hands-on patterns for building agents on top of existing model APIs.

• Track 15 lesson 14 (Agentic LMs and tool use). The from-scratch / lab-POV companion for the same topic.

• Track 20 (AI Agents and Tool Use). The full track-level treatment of the agent topic, ten lessons end to end. The right destination for a learner who wants more than the production-shipping primer this lesson provides.