Senior AI Engineer

What you'll do Define our evaluation practice: how we determine whether one model, prompt or system genuinely outperforms another, drawing on methods such as golden datasets, deterministic checks, reference-based metrics, LLM-as-a-judge and human evaluation. This is central to the role. Establish our MLOps foundations and optimize how we serve our on-prem, open-source LLMs, working closely with Platform Engineering: deployment pipelines, model and prompt versioning, evaluation harnesses, monitoring (latency, accuracy, drift), reproducibility, and the shared tooling and standards that help good practice spread across teams. Build production-grade LLM features hands-on, in close collaboration with Product Managers, Designers, Tech Leads and Software Engineers. Mentor our AI Engineers and set the patterns and best practices that define how we do AI across the team. Engage with subject-matter experts and customer-facing teams to ensure AI is used in ways that are meaningful, secure and scalable. What we're looking for (must-haves)   Several years of hands-on experience building and operating ML/LLM systems in production, ideally in a SaaS or product context. Several years working with evaluation and MLOps frameworks, enough to have real opinions about how to evaluate LLM systems: golden datasets, deterministic and reference-based tests, LLM-as-a-judge and its pitfalls, and regression analysis when comparing models. Solid MLOps practice: deployment, lifecycle management and monitoring in production. Deep, hands-on experience taking prompt engineering and RAG well beyond proof-of concept into systems that perform reliably in production. Hands-on experience with AI agents, including their failure modes and a realistic sense of when they're the right tool. Proficiency in Python (Java a plus). Strong interpersonal skills and a genuine mentoring mindset. You share direction and experience generously, and you lift the people around you without taking over. Comfort building foundations in a low-maturity environment. A strong commitment to data security, guardrails and responsible, ethical AI. Strong advantages   A track record of establishing ML/LLM practice from scratch (evaluation frameworks, testing discipline, MLOps foundations), rather than only working within practices someone else built. Ideally across more than one team or product. Hands-on experience running and serving models locally on your own GPU infrastructure, not just using open-source models through a hosted API. Experience fine-tuning open-source LLMs to make them more domain-specific. Experience translating emerging research into practical applications. What we offer   A foundation-owned company with a clear purpose to create a better world. Competitive salary. Support for your personal and professional development from day one, in an environment of professional exchange and learning among peers. A hybrid working model balancing on-site presence with remote flexibility. Health insurance and pension scheme. A sixth week of holiday. A strong social community with regular social events in a dynamic environment. An attractive lunch scheme in perhaps Aarhus' best canteen. How to apply Apply online with your CV and a cover letter. We expect you to understand, speak and write Danish (minimum CEFR B1, preferably B2 or higher) and to work on-site in Aarhus at least 3 days a week. We'll be conducting interviews in the second half of August and through September, so please send your application before then to be sure it's considered. Reach out to our Head of Software Engineering, Rasmus Hallenberg-Larsen, at rah@ramboll.com if you'd like to know more about the role