YouTube Channels¶
Channels that publish consistently high-quality LLM engineering content.
Technical depth¶
Andrej Karpathy Builds neural networks from scratch, including GPT. Essential for understanding transformers at the implementation level. - Let's build GPT: from scratch, in code, spelled out — the best transformer explanation that exists - Neural Networks: Zero to Hero — full series from micrograd to GPT-2
Yannic Kilcher Paper walkthroughs with detailed explanations. Covers attention, RLHF, constitutional AI, agents. - Good for: understanding research papers before implementing them
Practical engineering¶
Sam Witteveen Practical LangChain, LangGraph, RAG, and agent tutorials. Stays current with library changes.
James Briggs RAG pipelines, vector databases, LLM API integration. Detailed coding walkthroughs.
AI Jason LangChain, agents, production patterns. Good for intermediate practitioners.
Research and industry¶
AI Explained Clear explanations of new models, benchmarks, and research papers. Non-technical framing.
Two Minute Papers Quick summaries of AI research papers. Good for staying current without reading full papers.
Playlists to bookmark¶
- Andrej Karpathy — Neural Networks: Zero to Hero (full series)
- DeepLearning.AI YouTube — free content from their short courses
- Weights & Biases — MLOps and LLM engineering talks from practitioners
How to use YouTube for learning LLM engineering
Watch the first 10 minutes of a video. If the presenter is building something you need, watch the rest. If they're explaining something conceptual, pause and implement it yourself after each major point. Passive watching is not learning.