School S11 · F02

LLM Engineering

Serve, evaluate and specialize large language models.

For engineers who need to go below the API — quantization, serving, distillation, fine-tuning and building custom model stacks.

Learning outcomes

What graduates can do.

  • 01Serve open models at production latency and cost
  • 02Fine-tune and align models to specific domains
  • 03Distill and quantize for edge deployment
  • 04Build in-house model gateways and routers

Curriculum

9 courses.

Delivered as a portfolio-based sequence. Every course culminates in shipped work reviewed by working practitioners.

  1. 01Transformer Architecture
  2. 02Tokenization, Context & Attention
  3. 03Inference Servers (vLLM, TGI, TensorRT-LLM)
  4. 04Quantization & Distillation
  5. 05SFT, DPO & RLHF
  6. 06Evals for LLMs
  7. 07Model Routing & Gateways
  8. 08Self-Hosting Open Models
  9. 09Capstone: Ship a Custom LLM Stack

Tools & technologies

vLLMTGIHugging Facellama.cppPyTorchOllama

Career tracks

  • LLM Engineer
  • Inference Engineer
  • Foundation Model Team Engineer