Airtribe
Events
Upskill for free
Quiz zoneChallengesPractice with AIResources
ReviewsJob boardBusiness
AI for Builders

How LLMs Actually Work

Start here. Before agents, cost tradeoffs, or evaluation frameworks make sense, you need a mental model of what large language models actually do — how they're trained, how they generate text, and where they break down.

3 parts · ~25 min total

  1. 01The ML Concepts Behind LLMsThe vocabulary and mental models from machine learning that every builder needs — supervised learning, overfitting, fine-tuning, embeddings, and more.~8 min3 interactive demos→
  2. 02How LLMs Actually WorkFrom training to inference, next-token prediction to model selection — the foundational concepts every builder needs before diving into AI product work.~8 min3 interactive demos→
  3. 03Tokenization Up CloseThe model doesn't see words — it sees integers. Understanding tokenization explains half the mysteries builders encounter, from character counting failures to cost surprises.~9 min2 interactive demos→
Next seriesContext Windows & Memory