Course
AI Systems Engineering: Run, Operate & Evaluate
Make a model fit, serve it fast, keep it working, and prove it improved.
L4 · ExpertFast-movingKnown~3 h
What you’ll learn
- Explain why AI runs on GPUs and where the bottlenecks are
- Estimate a model's inference memory and how quantization changes it
- Describe serving throughput vs. latency and continuous batching
- Outline the MLOps/LLMOps lifecycle and what drift is
- Measure a system with precision, recall, and F1 and evaluate generation
- Name the main LLM security risk classes and their defenses
Prerequisites
llm-foundations
Module 1. Running Models
The hardware underneath, how much memory a model needs, and how it is served fast.
Module 2. Operating & Evaluating
Keep a model working in production, measure whether it improved, and defend it.