All cursus
AIintermediate
🧬How LLMs work: from scratch
Ten lessons that build up a modern LLM end to end — neural nets and backprop, attention and transformers, tokenization, pretraining, modern architectures and MoE, scaling laws, post-training (SFT/RLHF/DPO), and inference internals like the KV cache.
0 of 10 lessons complete
Sign in to track progress and earn a certificate.
Lessons in order
- 1AINeural Networks and BackpropagationStart
- 2AIAttention and TransformersStart
- 3AITraining: Optimization and RegularizationStart
- 4AILLM Tokenization in Depth: BPE, Byte-Level BPE, and SentencePieceStart
- 5AILLM Pretraining: Data, Loss, and What Actually HappensStart
- 6AIModern LLM Architectures: From Decoder-Only to Mixture-of-ExpertsStart
- 7AILLM Scaling Laws: From Kaplan to Chinchilla and BeyondStart
- 8AILLM Post-Training: SFT, RLHF, DPO, and Modern Alignment RecipesStart
- 9AILLM Inference Internals: KV Cache, Sampling, and Serving at ScaleStart
- 10AIContext windows: tokens, limits, and "lost in the middle"Start