A 16-course program on how LLMs actually work under the hood - tokenization, transformers, attention - plus hands-on practice with PyTorch, Hugging Face, and LangChain. A lot of the RAG background behind TalentLens came out of this.

Went back to fundamentals with this one - hooks, rendering behavior, component design, instead of just picking things up piecemeal on the job.

Same course series, but for Node - the event loop, async I/O, and what's actually happening under Express instead of just using it.