Level 3 · Advanced Under the Hood¶
Goal: understand what's inside the models you've been calling — transformers, tokenizers, embeddings — and learn to run and adapt open-weight models on your own hardware.
Modules¶
- Transformer Internals
- Tokenizers Deep Dive
- Embeddings & Semantic Search
- Running Local Models with Ollama
- Serving Open Models with vLLM
- Fine-Tuning Fundamentals (LoRA/QLoRA)
- Quantization & Inference Optimization
- Guardrails & Content Moderation
- Observability & Tracing
- Project — Self-Hosted LLM Stack
By the end of this level you'll be able to run, serve, and fine-tune open-weight models and instrument LLM applications end to end.