# Handling 1000's of alerts using LLMs

- Institution: CERN
- Speaker: Panagiotis Gkonis
- Year: 2026
- Language: English
- Source: [CERN](https://videos.cern.ch/record/3025724)
- Licence: [CERN](https://copyright.web.cern)
- Who can watch: Teens profiles and older (sources give no ratings; placed by Invisible Network from the source's own words)

Babysitting a handful of servers can be fun. But doing that with 5000? Nope. Who wants to deal with a daily stream of software and hardware alarms?

So what do you do? You automate, of course. And a year later you realize your automation code has gotten quite complex, despite not covering all the cases you've spotted. You just need to extend it further to handle more alarms, more faults, more edge cases. But you know there's no end to this and it will inevitably turn the code into a mess.

So what do you do? Naturally, the though crosses your mind: what if you used an LLM ?

00:00:00 Slide 1

00:00:26 Slide 2

00:00:56 Slide 3

00:01:09 Slide 4

00:01:24 Slide 5

00:01:44 Slide 6

00:01:58 Slide 7

00:02:26 Slide 8

00:02:57 Slide 9

00:03:48 Slide 10

00:04:27 Slide 11

00:04:53 Slide 12

00:06:05 Slide 13

00:06:29 Slide 14

00:07:42 Slide 15

## Lectures

1. [Handling 1000's of alerts using LLMs](https://invisiblenetwork.ai/learn/cern/3025724/0bfe29142531), 9 min (CERN's player)