LLM Inference
Learn how LLM inference works: foundations, GPU hardware, optimization techniques, and serving engines — from first principles to production-scale systems.
Foundations & Learning Path
Start with the 2026 learning path, then the foundational resources and the GPU hardware primer: what inference runs on and what it actually costs.
Optimization & Serving
Go from theory to throughput: the core inference optimization techniques, then how the leading serving engines — vLLM, BentoML, SGLang — compare in practice.
Projects & Deep Dives
Apply it: infrastructure projects worth building, Tiny-LLM on Apple Silicon, the official SGLang course, and the Mooncake KV-cache-centric serving architecture.
Frequently asked questions
What does the LLM inference category cover?
How LLMs are served in production: inference infrastructure, scaling and cost control, written from a DevOps perspective.
Is it beginner-friendly?
Concepts are explained before the infrastructure depth, so you can read for understanding or for deployment detail.
Are there hands-on examples?
Yes — guides include configuration and code examples you can adapt to your own stack.