Stenobird

Episode

HN819: Recipes for Automation – A Look Inside Eric Chou’s AI Networking Cookbook

Podcast
Heavy Networking
Published
Mar 20, 2026
Duration seconds
3709
Processing state
processed
Canonical source
https://packetpushers.net/podcasts/heavy-networking/hn819-recipes-for-automation-a-look-inside-eric-chous-ai-networking-cookbook/
Audio
https://feeds.packetpushers.net/link/12486/17302798/HN819.mp3
JSON
/v1/public/podcasts/heavy-networking/episodes/hn819-recipes-for-automation-a-look-inside-eric-chou-s-ai-networking-cookbook
Markdown
/podcast/heavy-networking/hn819-recipes-for-automation-a-look-inside-eric-chou-s-ai-networking-cookbook.md

Actions

  • POST https://stenobird.com/v1/public/podcasts/heavy-networking/episodes/hn819-recipes-for-automation-a-look-inside-eric-chou-s-ai-networking-cookbook/transcription-requests
    Idempotently request low-priority transcript generation for this episode.
  • GET https://stenobird.com/podcast/heavy-networking/hn819-recipes-for-automation-a-look-inside-eric-chou-s-ai-networking-cookbook.md
    Read the agent-friendly Markdown representation of this episode resource.

Summary

Network engineer Eric Chou introduces a systematic approach to integrating Large Language Models into network automation workflows. The discussion covers everything from setting up local models to using advanced prompting techniques for complex configuration tasks.

Topics

  • Network Automation
  • Large Language Models
  • Python Networking
  • Prompt Engineering
  • AI Security
  • Agentic Workflows
  • Infrastructure as Code
  • LLM Implementation

Highlights

  • Main idea: The AI Networking Cookbook provides a structured roadmap for moving from basic ChatGPT usage to building custom AI-driven automation tools
  • Practical takeaway: Effective prompting requires both 'persona' (telling the AI who it is) and 'technical context' (providing specific configuration examples)
  • Practical takeaway: Using 'agentic' coding assistants like Cursor can automatically ingest local project context, such as Python virtual environments and coding standards
  • Failure mode: Relying on the cheapest or default models often leads to poor reasoning and hallucinations; selecting the right model for the specific task is critical
  • Security takeaway: When using public LLMs, avoid uploading sensitive router configurations and utilize 'opt-out' settings in pro versions to protect data privacy

Chapters

  1. 1:00 Introduction to the AI Networking Cookbook: Eric Chou discusses the motivation behind his new book and its systematic approach to teaching AI for network engineers.
  2. 10:00 Foundations of AI for Automation: A look at the basics of LLM parameters, setting up local models, and the utility of using APIs versus GUIs.
  3. 24:25 The Power of Context and Vector Databases: How providing technical context and using vector databases improves the accuracy of LLM responses in networking tasks.
  4. 28:50 Crafting High-Quality Prompts: The importance of combining persona, environment details, and specific examples to get actionable automation code.
  5. 33:20 Agentic Workflows and Coding Assistants: How modern tools like Cursor use local file context to understand Python environments and coding standards automatically.
  6. 38:05 Model Selection and Hallucination Risks: Discussing the trade-offs between model cost, performance, and the risks of using inadequate models for complex logic.
  7. 56:55 Security and Privacy in AI Networking: Best practices for using public LLMs without compromising network security or leaking sensitive configuration data.