Job Description

Do you want a job with a purpose?

And do you want to make healthcare safer, better and more reliable?

Join our Team!

Join us as our AI Engineer (m/f/d) - Healthcare IT at Dedalus, one of the World’s leading healthcare technology companies, in Graz to do the best work of your career and make a profound impact in providing better care for a healthier planet.

You will contribute to the engineering behind our agentic AI roadmap: designing, building, configuring, evaluating, and operating production-ready agents that work reliably and safely in a regulated healthcare environment. This is an application- and systems-layer engineering role. Your success is measured by delivering agents that survive contact with real users and real clinical workflows.

You will work closely with clinical, product, and domain experts, but your core craft is shipping and qualifying agentic systems, building the evaluation infrastructure, and optimizing the cost and latency of running them at scale.

What you'll do

• Build or configure production-ready agents. Design and ship multi-step, tool-using agents with memory and state for clinical and operational workflows. Own them end-to-end, including orchestration, and integration. Support our delivery and operation teams with deployment, and finding and fixing the root causes of failures.

• Own the evaluation harness. Build and maintain offline eval datasets, combine LLM-as-judge with deterministic checks, and wire regression gates into CI so quality is enforced before release — not discovered in production.

• Design and optimise RAG. Build retrieval pipelines over structured and unstructured clinical text: chunking, embedding and retrieval strategy, reranking, and grounded/cited generation. Measure retrieval quality (faithfulness, context relevance, recall), not just end-output vibes.

• Qualify agents for a regulated setting. Define acceptance criteria and failure-mode analyses; design guardrails, human-in-the-loop, and escalation paths; and produce the validation and verification evidence and traceability expected under Medical Device Regulation (MDR) and the EU AI Act for clinical AI.

• Instrument and observe in production. Enable tracing every step an agent takes — LLM calls, tool calls, retrieval steps — monitor for quality drift, and close the loop by turning production failures into new eval cases.

• Optimise cost and latency. Treat cost-per-task as a first-class metric. Apply model routing/selection, caching, token budgeting, batching, and prompt efficiency.

Essential Requirements

• Demonstrated, shipped production AI systems (preferably LLM/agent systems) — real systems that ran in front of real users, not notebooks or proofs-of-concept. Be ready to walk us through one, including how it failed and what you did about it.

• Strong software engineering fundamentals: Python, API design, automated testing, version control, containerisation.

• Hands-on agent orchestration with at least one framework and a clear understanding of how agentic systems break in production.

• Practical RAG experience: embeddings and vector stores, retrieval design, and retrieval-specific evaluation.

• LLM evaluation experience: building or operating eval harnesses and observability/tracing tooling.

• Comfort with ambiguity and unstructured problems, plus clear communication — able to explain technical trade-offs to clinical and product stakeholders.

• A degree in computer science, software engineering, or a related field.

• Proficiency in English and German language.

• Note on experience: we are looking for years of strong software engineering plus recent, hands-on agent work Agent engineering is a young discipline, and we are not expecting a long tenure in it.

Desirable Requirements

• Healthcare/clinical-software and data exposure.

• Strong foundation in statistics and probability theory.

• Awareness of MDR and EU AI Act implications for clinical AI (high-risk classification, human oversight, data governance, post-market monitoring).

• Deep cloud platform knowledge (especially AWS) and infrastructure-as-code.

• Model adaptation or fine-tuning.

• Familiarity with guardrail and safety frameworks for LLM applications.

• Proficiency in Italian language.

We are Dedalus, come join us

Dedalus is committed to providing an engaging, rewarding work experience that reflects the passion our employees bring to our mission of helping clinicians and nurses deliver better care to their served communities.

Our company fosters a culture where employees are encouraged to learn and innovate, and to enable and enhance clinical co-operation and processes while making a meaningful difference for millions of people around the world. Each person is the end point and the starting point of the Group’s activities and the ultimate beneficiary. For this reason, we are so proud of doing our very special jobs each day.

Our company is enriched by a diverse population of 7,600 people in more than 40 countries that work together to innovate and drive better healthcare options for millions of patients around the world. We are the people of Dedalus.

Application closing date: 31st of July 2026

Our Diversity & Inclusion Commitment sets out Dedalus’ approach to ensuring respect, inclusion and success for all our colleagues and the wider communities we operate in. It is imperative for us to share our commitment and dedication to ensure an inclusive and diverse workplace. We recognise that we have improvements to make, and, on this journey, we must remain authentic and realistic but also ambitious.

Our diversity & inclusion commitment – Dedalus Global

Life Flows Through Our Software

#LI-PA1