AI Software Engineer | Spain

Description

You will build AI systems that run in live enterprise settings, not demos or pilots that never reach production. The role suits someone who has shipped a multi-agent system, owned an evaluation harness, and dealt with production failures firsthand.

The role is an AI Engineer position with agentic and applied focus, based in Spain. The company is a global professional services firm working in more than 120 countries. Demand for AI engineers is rising faster than the supply.

Responsibilities

• Work inside client engineering teams to design, prototype and ship agentic solutions through workshops, proofs of concept, pair-coding sessions and architecture reviews.
• Design and build production agentic systems from start to finish, covering multi-agent orchestration, retrieval-augmented generation (RAG) pipelines, policy-based routing, tool calls, memory handling and lifecycle observability.
• Own RAG pipelines, including embeddings, chunking, vector search and context window tuning against measured quality targets.
• Connect several LLM providers (OpenAI, Anthropic, Vertex AI and open-source models) behind one abstraction, with fallback routing and control over tokens, cost and latency.
• Run LLMOps in production: evaluation harnesses with real quality metrics, prompt versioning, observability tooling such as LangSmith or Braintrust, and cost and safety monitoring.
• Measure agent accuracy, latency, safety and cost-effectiveness, then explain findings and recommendations to client stakeholders in business terms.
• Build reusable patterns, accelerators and playbooks that help later client engagements start faster.

Requirements

• Solid software engineering background in production environments.
• Hands-on experience designing and deploying agentic AI in production. This is essential.
• Production-depth experience with orchestration frameworks such as LangGraph, CrewAI, AutoGen or equivalent, not tutorial-level work.
• Experience calling LLM APIs (OpenAI, Anthropic, Vertex AI) in production code, including provider abstraction and trade-offs between tokens, latency and cost.
• Ownership of RAG pipelines: embeddings, chunking, vector databases and context engineering.
• Working knowledge of LLMOps: evaluation harness design, prompt versioning and production observability.
• Cloud-native skills with Kubernetes, Docker, microservices, serverless, CI/CD and infrastructure as code (Terraform or Helm).
• Strong Python. Java or an equivalent backend language is acceptable. Experience debugging and observing production systems.
• Quality of experience counts more than years. A candidate who has shipped three production agentic systems in four years is preferred over a generalist with passive AI exposure.

Conditions

Based in Spain.

Hiring decisions are made regardless of age, race, creed, colour, religion, sex, national origin, ancestry, disability, veteran status, sexual orientation, gender identity or expression, genetic information, marital status, citizenship, or any other characteristic protected by applicable law.

Benefits

• Vendor fellowship access inside the engineering teams of Anthropic, OpenAI, Microsoft and Google.
• A direct pathway into the Forward Deployed Engineer programme.