Stenobird

Episode

HN812: Nokia EDA: AI Ops You Can Trust (Sponsored)

Podcast
Heavy Networking
Published
Jan 30, 2026
Duration seconds
3061
Processing state
processed
Canonical source
https://packetpushers.net/podcasts/heavy-networking/hn812-nokia-eda-ai-ops-you-can-trust-sponsored/
Audio
https://feeds.packetpushers.net/link/12486/17266989/HN812.mp3
JSON
/v1/public/podcasts/heavy-networking/episodes/hn812-nokia-eda-ai-ops-you-can-trust-sponsored
Markdown
/podcast/heavy-networking/hn812-nokia-eda-ai-ops-you-can-trust-sponsored.md

Actions

  • POST https://stenobird.com/v1/public/podcasts/heavy-networking/episodes/hn812-nokia-eda-ai-ops-you-can-trust-sponsored/transcription-requests
    Idempotently request low-priority transcript generation for this episode.
  • GET https://stenobird.com/podcast/heavy-networking/hn812-nokia-eda-ai-ops-you-can-trust-sponsored.md
    Read the agent-friendly Markdown representation of this episode resource.

Summary

Moving beyond AI hype requires a foundation of structured data and multi-vendor abstractions. This episode explores how Nokia's EDA platform uses YANG schemas and event-driven architecture to make AI operations deterministic and trustworthy.

Topics

  • AI Ops
  • Network Automation
  • Event-Driven Automation
  • YANG Models
  • Multi-vendor Networking
  • Large Language Models
  • Network Observability
  • Infrastructure as Code

Highlights

  • Main idea: AI in networking fails without a foundation of well-defined schemas and structured telemetry
  • Practical takeaway: Use YANG models and network-wide abstractions to provide LLMs with the necessary context for accurate reasoning
  • Failure mode: Relying on non-deterministic LLM outputs without enforcing constraints through automation primitives can lead to dangerous hallucinations
  • Main idea: An effective AI Ops platform should treat AI agents as interfaces that interact with existing, repeatable automation workflows
  • Practical takeaway: Implement an app-based architecture to allow for modular upgrades and extensible network capabilities without system-wide downtime

Chapters

  1. 1:00 Introduction to Nokia EDA: An overview of Nokia's Event Driven Automation (EDA) platform and its focus on multi-vendor network infrastructure.
  2. 4:35 The Power of Abstraction: How using primitives and YANG schemas allows for consistent configuration and status monitoring across different vendors.
  3. 8:55 Providing Context to LLMs: The importance of presenting Large Language Models with structured health scores and fabric definitions to ensure accurate reasoning.
  4. 12:35 Network-Wide Abstractions: Moving from interface-level configuration to higher-level abstractions like fabrics and services.
  5. 16:20 Extensible Data Models: How engineers can implement custom data models and syntax within the automation framework.
  6. 20:25 AI Agents and Automation Tools: Defining the relationship between AI chatbots as agents and existing automation workflows as the underlying tools.
  7. 24:10 Enforcing Determinism: Using schemas to constrain LLM outputs and prevent hallucinations in production environments.
  8. 28:15 Modular Platform Architecture: The benefits of an app-based model for upgrading network capabilities without disrupting the core foundation.