DevOpsDays BCN Meetup - October 2026
Details
Hello, fellow DevOps enthusiasts 🤓
On October 1st, we'll meet again for a couple of insightful talks, hosted by Allianz Technology Spain!
📍Location: Avda Icà ria 199, Barcelona
------
đź“… Agenda
* 18:30 - 18:40 Welcome from the hosts
* 18:45 - 19:30 Nobody's on fire, everybody's tired
* 19:30 - 20:15 TBC
* 20:15 - 21:00 Networking & Snacks
------
🎤 Talks & Speakers
- Title: Nobody's on fire, everybody's tired
- Speakers:
David Castañeda - Senior DevOps Engineer at Xcelirate
Eduard Epure - Security Engineer at Mews - Duration: 30-45 mins
- Abstract: One of us barely reads the AI's code. Neither of us has caused an incident. Both of us are exhausted. David runs user-facing infrastructure where one bad change is visible to customers; he reads every line the AI writes and pushes back on plenty of it. Edu builds an internal platform used by a couple hundred engineers; he ships AI-generated changes once a pipeline of syntax checks, security checks, and an AI reviewer goes green, and spends his attention running specs in parallel. Two senior engineers, two companies, two opposite answers to the same question: how much should you trust code you didn't write? We'll each defend our way of working, then tell you what we found when we compared notes.
------
- Title: How to deploy Agentic AI with GitOps by reusing Infrastructure as Code
- Speaker: Pere Alcoberro Turu - Head of Platform Engineering at Allianz Technology
- Duration: 45 mins
- Abstract: AI agents are becoming another production workload, but deploying them should not require inventing a completely new delivery model. In this talk, I’ll show how we can apply the same Platform Engineering and GitOps principles we already use for traditional applications to Agentic AI.
Using AWS Bedrock AgentCore, Crossplane and ArgoCD, I’ll walk through how I approached agent deployment as a reusable platform capability. The key idea was not to start from scratch: I reused patterns from existing Crossplane compositions for ECS-based applications and adapted them to the requirements of AgentCore.
I’ll cover:
- Deploying AI agents to AWS Bedrock AgentCore through Crossplane compositions.
- How existing Infrastructure as Code patterns from ECS-based applications can be reused and evolved instead of building a separate AI deployment stack.
- Modeling the agent runtime, networking, IAM and supporting AWS resources behind a standardized Crossplane API.
- Using Git as the source of truth and ArgoCD as the reconciliation layer, so an agent can be deployed and managed through the same GitOps workflow as an application.
- Lessons learned from moving from CLI/CDK and proof-of-concept deployments toward a repeatable, declarative approach.
The goal is to demonstrate that Agentic AI does not need a special deployment process. By treating agents as applications and exposing their infrastructure through reusable platform APIs, teams can deploy them in a standardized, auditable and repeatable way—with a pull request becoming the interface to the deployment platform.
------
🔑 Important Access Note
Our host requires attendees to provide their ID to grant us access to their offices. Do not forget to fill in the "NIF - NIE - Passport" field when registering!
