Building with LLM APIs

From one API call to a feature you can ship: a prompt kept as code, output your program can trust, answers that stream into the browser, failures handled on purpose, a bill with a ceiling, tests that do not flake and a safe way to change the model.

Watch lesson 01

Free preview01 What one call sends and returns

Lessons
18
Modules
3
Running time
~3.3 hours
Before you start
Read TypeScript and JSON
Free to watch
6 lessons
Last updated
Nov 7, 2026

Part of

[01] Access

[02] Who it's for

  • Engineers adding an LLM feature to a real product, not building a chatbot or training models.
  • Developers who have called a model API from a tutorial and want to know what it takes to ship one.
  • No machine learning background is needed. You should know HTTP, a web server and unit tests.

[03] What you'll learn

  • 01Read any LLM API request and response and say what every part does, usage and stop reason included
  • 02Keep a prompt in your code with a version, and keep the user's text from acting as instructions
  • 03Get output your program can trust: a schema, a stop-reason check and checks on the meaning
  • 04Stream answers into a browser through a real server and proxy, with a working Cancel button
  • 05Decide for every failure whether to retry, re-ask, degrade or stop, and put a ceiling on the bill
  • 06Test a model feature without flaky tests, ship changes behind a flag and move to a new model safely

[04] Syllabus

Module 01

The call

6 lessons68 min
  1. 01
    10m
  2. 02
    8m
  3. 03
    11m
  4. 04
    10m
  5. 05
    12m
  6. 06
    13m
  7. Module quiz

    5 questions, about 5 minutes

Module 02

The feature

6 lessons77 min
  1. 07
    13m
  2. 08
    12m
  3. 09
    12m
  4. 10
    13m
  5. 11
    12m
  6. 12
    10m
  7. Module quiz

    5 questions, about 5 minutes

Module 03

Production

6 lessons67 min
  1. 13
    12m
  2. 14
    9m
  3. 15
    11m
  4. 16
    7m
  5. 17
    11m
  6. 18
    11m
  7. Module quiz

    5 questions, about 5 minutes

[05] How it's made

The curriculum is curated by karnstack and reviewed by senior engineers in the industry before it ships. Narration is an AI voice (ElevenLabs) reading human-written, human-reviewed scripts. Read how courses are made.