In September 2025, I spoke at RONOG 10 (Romanian Network Operators Group), presenting a practical introduction to AI for network operators and engineers — from simple automation scripts to intelligent agents that can reason and interact with infrastructure.
The talk, titled "From Hacker Scripts to Smart Agents: A Gentle Introduction to Real-World AI," demonstrated how large language models are transforming network operations — not through hype, but through practical tools that integrate with real infrastructure.
About RONOG
RONOG (Romanian Network Operators Group) is an annual conference bringing together network operators, ISPs, technology professionals, and government representatives. It provides a platform for sharing technical knowledge about internet infrastructure, cybersecurity, and emerging technologies in an informal, focused environment.
The Talk: From Scripts to Agents
The presentation walked through a progression familiar to most network engineers:
- Starting with simple command-line scripts for automation
- Moving to tools that understand natural language prompts
- Advancing to intelligent agents that can reason, plan, and adapt
- Integrating custom tools using Model Context Protocol (MCP)
The focus was on real-world application: how to integrate network devices and servers into a central AI system using custom MCP tools. I worked with the Interlan team to validate ten practical scenarios that would produce actual value in a real network operations business.
"No buzzwords, just real use cases — showing how AI tools can assist engineers, automate workflows, and adapt to unexpected input."
Watch the Full Presentation
Making AI Accessible to Network Operators
The session was designed for technical audiences without prior AI experience. By starting from familiar territory — scripts, prompts, CLI tools — and gradually moving toward more advanced systems like agentic AI, the talk made these concepts approachable and actionable.
The response from the network operations community validated the approach: automation is the future, but it works best when built on tools that engineers actually understand and can integrate into their existing workflows.
This RONOG 10 presentation demonstrated that AI in network operations isn't about replacing engineers — it's about giving them better tools to solve real problems faster and more reliably.