This skill is about building software on top of a large language model (LLM) — a model trained on text that generates a response one piece at a time. It is not about training models. Almost everything you control lives in one place: what you put in front of the model, and what you do with what comes back. The recurring surprise is that the model is a stateless, probabilistic component in an otherwise ordinary system, and most bugs come from treating it as either a database or a person. Read this for the shape; the practice questions drill the edges.
Each chapter opens with the short version. Tap one to read the detail.
A. What the model is actually doing when it answers
~2 min
The model predicts a likely continuation, one token at a time. Nothing in that process distinguishes a remembered fact from a fluent guess, which is why hallucination is a property of the mechanism rather than a bug in a particular answer.
B. Tokens, tokenization and the context window
~2 min
The model reads and writes tokens, not characters, and the context window is a working budget that everything shares — system prompt, history, tool definitions, documents, and the response itself. More context is not automatically better.
C. Sampling controls and determinism
~2 min
Output is drawn from a probability distribution, and temperature controls how adventurous that draw is. Lowering temperature to zero makes output more focused but does not make it reproducible.
D. The Messages API surface
~2 min
A request is a list of alternating user and assistant turns plus a separate system parameter. The API is stateless: the conversation exists only because you resend it every time.
F. Prompt caching
~2 min
Prompt caching stores a prefix of your request so repeated calls skip reprocessing it. It only pays off when the cached part is genuinely identical, byte for byte, across requests.
G. Prompt engineering
~2 min
Prompt engineering is optimization, and optimization without measurement is guessing. Define what success means and build a way to test it before you start rewriting wording.
I. Tool use and function calling
~2 min
Tool use does not let the model run code. The model emits a structured request naming a tool and its arguments; your application executes it and sends the result back. Everything downstream follows from that split.
J. Agents and the agent loop
~2 min
An agent is the tool loop run to completion: the model plans, calls tools, reads results, and decides when it is done. The hard parts are not the model — they are the loop's stopping conditions, permissions, and context growth.
L. Retrieval-augmented generation
~2 min
Retrieval-augmented generation means fetching relevant material at request time and putting it in the context window. The model does not reach out and retrieve anything — your code does, before it asks.
M. Embeddings, vector search and reranking
~2 min
An embedding turns text into a vector so that similar meanings sit close together, which is what makes semantic search possible. The Claude API does not include an embedding model, so picking one is its own vendor decision.
N. Model Context Protocol
~2 min
The Model Context Protocol (MCP) is an open standard for connecting models to external tools and data, so an integration written once can be reused across applications instead of rebuilt per vendor.
O. Evaluation
~2 min
An evaluation is a repeatable test of model output against your success criteria. Volume with automated grading beats a handful of carefully hand-read examples, because you cannot iterate on what you cannot re-run.
P. Prompt injection, jailbreaks and output safety
~2 min
There are two distinct threats. In one the user is the adversary; in the other the user is fine and the danger arrives inside content the model reads on their behalf. The second is the one that surprises people building agents.
S. Reliability and production hygiene
~2 min
Limits are enforced per minute on requests, input tokens, and output tokens. Exceeding one returns an error carrying a wait time — and honouring that number is most of what production reliability requires.
Written by Keentune. We are not affiliated with or endorsed by the organizations whose documentation informs this guide, and any linked sources belong to their respective owners.
All exam, test, and product names and trademarks are the property of their respective owners and are used here for identification and reference only. Keentune is independent study practice — not affiliated with, authorized, or endorsed by any of these organizations.