AI Engineer

Location: Madrid, Spain (Remote — with regular in-person collaboration at a local WeWork alongside our Madrid tech team)Media Radar, now including the data and capabilities of Vivvix, powers the mission-critical marketing and sales decisions that drive competitive advantage. Our competitive advertising intelligence platform enables clients to achieve peak performance with always-on data and insights that span the media, creative, and business strategies of five million brands across 30+ media channels. By bringing the advertising past, present, and future into focus, our clients rapidly act on the competitive moves and emerging advertising trends impacting their business. We are building a best-in-class AI team focused on delivering advanced capabilities that empower our data organization. This is not a role for theoretical research; it is for an AI Engineer with a strong Software Engineering backbone. You will be expected to be highly independent, taking ownership of problems from end-to-end. We value "Modern Productivity"—you should leverage LLMs and AI coding assistants to bypass boilerplate and ship code faster. In our view, manual line-by-line coding is outdated; we want engineers who use every tool at their disposal to increase output without sacrificing reasoning or quality. Stack Highlights: Core: Expert Python (Non-negotiable), PostgreSQL + pgvector, SQLAlchemy. AI/LLM: Azure OpenAI, Lang Chain, Langfuse, Multi-agent orchestration. Productivity: ChatGPT/Copilot for development, Docker, async I/O.

Key Responsibilities: Independent System Building: Own the design and implementation of scalable ML solutions, moving rapidly from problem statement to working code. Classification & Attribution: Build and fine-tune systems for complex data, such as identifying ad creatives or global brand deduplication by cross-referencing government registries. Advanced Reasoning: Implement "Chain of Thought" prompting and multi-agent workflows (e.g., classification agents vs. reasoning agents) to solve complex attribution tasks. Data Engineering Rigor: Optimize vectorization, manage migrations via Alembic, and ensure high-throughput ingestion with minimal supervision. Operational Excellence: Implement guardrails and observability (Langfuse) to ensure that high-velocity output remains explainable and robust.