---
title: "vLLM internals and production serving"
lesson_id: "09"
---

# vLLM internals and production serving

- **Lesson ID:** 09
- **Goal:** Serving is batching, KV cache, and tensor parallel. Read PagedAttention as an operator, then serve your small checkpoint.
- **Human lesson:** [09-vllm-internals-and-serving.html](09-vllm-internals-and-serving.html)
- **Phase:** Phase 5 Inference

## Prerequisites

- The previous week's artifact, or a written note if this is week 1.
- Python 3.11+ and a laptop. CUDA is useful from week 5 onward and not required to read.

## Inputs, outputs, and artifacts

- **Inputs:** The replica from the prior week.
- **Outputs:** Core project 7 starts: an OpenAI-compatible vLLM server and a latency log.
- **Artifacts:** Lesson notes plus the code named in the human page.

## Agent build steps

1. State the week 1 principle: you do not need to pretrain an 800B model to understand the shape of the problem; you do need progressively realistic replicas.
2. Teach the topics on the human page without inventing paper URLs or news claims.
3. Keep starter code in fenced blocks that match the human lesson.
4. End with the assignment and the opinion checkpoint.
5. Use Edge FDE tone: ship, do not sightseeing. Opinion checkpoints matter.

## Constraints

Plain spoken English. No quizzes. No em dashes. No invented metrics, dates, or citations. Brand Edge FDE only. Name well-known papers by title only: Attention Is All You Need, InstructGPT, Direct Preference Optimization, ZeRO, PagedAttention/vLLM, FlashAttention, Scaling Laws for Neural Language Models.

## Key concepts

- PagedAttention treats KV cache like virtual memory pages.
- Prefill and decode have different bottlenecks.
- Tensor parallel splits weights. Batching splits time.
- Throughput and latency are a pair you measure together.

## Takeaways

- Serve something and log TTFT plus inter-token time.
- Estimate KV bytes before you talk about GPU size.
- Name PagedAttention/vLLM as the serving replica, not as hype.

## Acceptance checks

- Human HTML has Key concepts and Takeaways.
- Opinion checkpoint is present.
- No em dash and no fake URL.
- [ ] Proceed to [lesson 10 brief](10-triton-and-hand-rolled-kernels.llms.md).