Senior AI Engineer (f/m/d)

Awin is part of the Axel Springer group. The AI Solutions team in the Growth Domain builds AI-native products used by advertisers, publishers, and partner managers globally to manage and grow affiliate programs. The team uses a Python-based stack on AWS and practices scrum.

• 8+ years of professional software engineering experience with a track record of shipping production systems.
• Strong Python and backend system design skills.
• Hands-on experience building and running production LLM-based systems.
• Deep experience with agent-based architectures using frameworks like LangGraph or Deep Agents.
• Strong knowledge of retrieval and context construction (RAG, embeddings, chunking, ranking).
• Experience evaluating and debugging AI systems with structured evals, regression tests, and observability tools such as LangSmith.
• Familiarity with emerging agent orchestration standards (MCP or similar).
• Experience with vector databases and hybrid retrieval.
• Understanding of LLM failure modes like prompt injection, content safety, context leakage, and mitigation techniques.
• Strong fundamentals in distributed systems, API design, and cloud-native architectures on AWS (ECS, Lambda, S3, API Gateway).
• Proficiency with PostgreSQL, Redis, Docker, and CI/CD pipelines like GitHub Actions.
• Collaborative engineer comfortable working in Scrum teams.

Nice to have:

• Experience with experimentation frameworks (A/B testing, offline-to-online metric correlation).
• Experience with cost-aware AI engineering at production scale, including monitoring token usage, model selection, caching, and routing strategies.

• Design, build and operate customer-facing AI features end-to-end, including conversational experiences and intelligent automation.
• Architect agent-based systems with multi-step reasoning, tool execution, sub-agent orchestration, state management, streaming, structured outputs, persistent memory, and human-in-the-loop approval flows.
• Decide how the system grounds, retrieves, and constructs context.
• Select and defend the appropriate pattern (retrieval, structured tools, agentic flows) for each problem.
• Treat prompt and context engineering as first-class engineering work.
• Own AI quality by building evaluation datasets, regression tests for prompts and agents, and debugging model outputs, retrieval, and tool calls.
• Make AI production-ready with observability, tracing, reliability, cost, latency, and safety controls against hallucination, prompt injection, and context leakage.
• Make and explain trade-offs between accuracy, latency, cost, and safety with engineering and product partners.
• Turn ambiguous product ideas into concrete technical designs and influence the roadmap.
• Mentor engineers, contribute to design reviews and hiring, and raise standards for production-ready AI.

Diversity & Inclusion are paramount; the company welcomes all backgrounds, identities, and experiences. Support is available during application or interview processes if needed.