Engineer - AI Platform

Ebury is building an Agentic Data Plane (ADP) to enable safe and scalable AI product deployment. The AI Platform Engineer will design core components like agent-to-agent communication, tool registry, and governed data access, while ensuring observability and integration with the data platform. This role is part of the Data Platform Engineering team in Madrid or Malaga, working in a hybrid model (4 days office, 1 day remote).

• Design and deliver core components of Ebury's Agentic Data Plane, including agent-to-agent (A2A) communication, the Agent Tool Registry, and governed data access layers.
• Build platform primitives for secure tool use coordination, including 'on behalf of' (OBO) patterns, centralized policy guardrails, and audit logging with full replayability.
• Provide end-to-end observability for agentic products: data dependencies, traces, evaluation signals, and accountability as a day-one requirement.
• Integrate the ADP with the Data Platform's governed foundations — catalog, metadata, data warehouse, and vector retrieval.
• Collaborate with domain teams to onboard pilot agents and help them adopt platform standards and runtime patterns.
• Drive innovation and explore emerging technologies (LLM evaluation, prompt safety, agent memory) to enhance AI platform capabilities.

• Self-motivation and curiosity, with a desire to learn and grow in a fast-paced environment.