Senior Engineer - Agentic Platform and AgentOps

• Flexible working hours and hybrid work (3 days in the office and 2 days remote).
• A flexible compensation package that includes: meal vouchers, childcare vouchers, medical insurance, life insurance, employee discounts, and more. And if you live in Madrid, a company shuttle bus is available to get you to the office.
• Access to an attractive pension plan.
• Summer intensive schedule.
• You'll be in charge of your own learning journey, with access to thousands of resources to keep advancing your technical expertise. Plus, you'll receive an additional €1,100 per year to further develop your skills.
• 25 vacation days per year (you would receive the proportional days starting from your start date).
• Additionally, at Mapfre we do not work on December 24, December 31, or January 5.
• 28 hours per year for personal matters + 3 recoverable days during Easter week.
• The opportunity to participate in volunteer projects.
• Free paddle tennis courts and gym.
• A great place to work! We're ranked among the 50 best companies to work for in Spain according to Forbes magazine.

We are diverse, and that makes us unique!

Give your career a fresh start and keep growing professionally at Mapfre Tech! You'll join an international environment where you can innovate by working on global projects focused on the latest technology trends. All of this in a flexible, agile setting surrounded by excellent professionals from around the world... Are you ready?

The purpose of the role is to ensure that our Agentic Platform evolves and operates in a reliable, secure, observable, and governed manner, industrializing the agent lifecycle through AgentOps, automation, and Production Ready criteria.

• Direct impact on a transversal platform for multiple countries, products, and teams.
• Real ownership over operational architecture, automation, technical governance, reliability, and roadmap.
• Cutting-edge challenges in multicloud agents, observability, AI security, evaluation, FinOps, and 24x7 operation.
• Opportunity to build the standards that will define how agents are brought into production.

We are looking for a very hands-on Senior Engineer who will technically lead the evolution and operation of the Agentic Platform. Their goal will be to take AI agents from experimentation to industrial, secure, and governed operation, applying AgentOps practices in the same way that DevOps industrializes software.

You will work on a multicloud and multi-framework ecosystem that may include Microsoft 365 agents, Copilot Studio, Azure AI Foundry, AWS Bedrock and AgentCore, GCP Vertex AI, LangGraph, ADK, Semantic Kernel, MCPs, containers, and shared corporate services. You will collaborate with Engineering Platform, Architecture, Security, AI Center, Operations, Observability, FinOps, and product teams.

Your mission will be to guarantee that the Agentic Platform is a reliable and scalable corporate capability, with a clear operating model, unique catalog and inventory, security controls, end-to-end observability, continuous evaluation, cost management, and secure mechanisms for isolation, rollback, and decommissioning.

• Assume technical reference and operational ownership of the Agentic Platform, defining priorities, standards, roadmap, and evolution model.
• Design and maintain the end-to-end AgentOps cycle: provisioning, validation, deployment, evaluation, publication, monitoring, change, suspension, reactivation, rollback, and decommissioning.
• Define and automate the Production Ready Gate so that no agent reaches production without an owner, criticality, technical identity, permissions, telemetry, SLOs, runbooks, continuity, and approval.
• Operate and evolve runtimes, LLM Gateway, MCP servers, catalogs, marketplace, identity, secrets, connectivity, containers, telemetry, and evaluation mechanisms.
• Build unified observability with metrics, logs, and traces for latency, errors, sessions, tool calls, retries, drift, quality, token consumption, quotas, and cost.
• Define alerts, dashboards, and runbooks for the Operations Center, including kill switch or safe-stop procedures with evidence preservation.
• Industrialize deployments through GitHub, CI/CD, infrastructure as code, automated testing, policy as code, and remediation automation.
• Integrate inventory and governance with OneInventory, Entra ID, Purview, Defender, Microsoft Graph, and equivalent services.
• Coordinate advanced troubleshooting and root cause analysis, converting incidents into preventive controls and reusable automations.
• Measure KPIs for availability, reliability, adoption, security, quality, cost, Production Ready, and operational efficiency.
• A platform ready to host and operate critical enterprise agents homogeneously.
• A reliable inventory with functional and technical owners, criticality, identity, dependencies, data, permissions, and environment.
• Sufficient telemetry to detect degradations and reconstruct relevant executions.
• Reduction of MTTR through runbooks, automated diagnosis, and controlled remediation.
• Control of consumption and costs by agent, model, provider, country, product, and use case.
• Auditable traceability of the lifecycle, changes, access, and actions on tools.

1. Discover and inventory: Agents, owners, identities, dependencies, data, tools, criticality, and environment.

2. Validate Production Ready: Architecture, security, telemetry, SLOs, runbooks, continuity, costs, and approvals.

3. Deploy with control: CI/CD, IaC, testing, versioning, environment segregation, and change evidence.

4. Observe and operate: Health, performance, quality, security, usage, and cost; response via runbooks.

5. Improve or retire: Remediation, drift correction, optimization, rollback, or traceable decommissioning.

Las Tablas, Madrid.