Senior Research Engineer

Senior Research Engineer - Applied LLM Systems (m/w/d)

Who We Are

Ascléa develops cutting-edge AI solutions for hospitals. We are an ISO 13485-certified medical device manufacturer with several years of experience in MDR-certified products for clinical decision support.

With our Clé platform, we provide hospital staff with secure and controlled access to language models, clinical knowledge, and AI-powered workflows. Building on this, we are gradually developing an agent-based cockpit for hospital operations.

What It's All About

As we develop and roll out Clé, we regularly encounter technical questions that require more than just a quick answer: Which approach works reliably? How do we measure this effectively? Is language model A actually better than B, or does it just appear that way in a benchmark? And when is a solution good enough to be deployed in production? You’ll tackle these issues largely on your own, addressing them methodically and with the necessary depth—from defining the problem precisely to selecting and investigating suitable approaches, all the way through to evaluation and experimentation. The end result will be robust findings, clear decisions, and implementations that the core team can build on directly.

In doing so, you’ll work closely with our Chief Product Officer and the development team. Within the scope of a given problem, you’ll have a great deal of freedom in how you approach it, which approaches you explore, and which ones you discard based on the results. We’ll decide together which results are ultimately incorporated into the product and in what form.

What You Do

• Structuring: You transform an open-ended technical question into a clearly defined problem and determine the criteria by which a good solution can be measured.
• Selecting Approaches: You familiarize yourself with the state of the art, identify viable solutions, and independently decide which ones to explore in greater depth.
• Evaluate: You develop evaluation methods for models, prompts, and pipelines, and critically assess whether datasets, metrics, and experimental setups actually answer the question we’re interested in.
• Implementation: You develop and test the most promising approaches and write code in a way that allows the core team to take it over and further develop it.
• Interpreting Results: You present your results in a way that enables an informed product decision to be made: What works, how robust is the result, what effort is involved, and what risks or open questions remain?

What You Bring to the Table

• You have been working independently on challenging technical problems for at least five years - as an engineer, data scientist, in applied research, or in a comparable role.
• You write clean, well-structured Python code that others can understand, adopt, and build upon.
• You have practical experience with large language models, particularly with their evaluation. You have developed your own tests and evaluation methods and understand how strongly results can depend on datasets, metrics, and experimental designs.
• You work methodically: formulating hypotheses, measuring, critically reviewing results, and documenting them in a transparent manner are second nature to you. A Ph.D. can be a good background for this role but is not a requirement. Comparable experience in independent technical or scientific work is equally relevant.
• You can confidently interpret quantitative results and assess how robust a comparison or evaluation actually is. When necessary, you can independently familiarize yourself with new statistical or methodological approaches.

Not every point applies to you? If your past work shows that you can independently analyze open technical problems and arrive at reliable solutions, we’d still love to receive your application.

How We Work

• Remote-first, with offices in Berlin-Mitte and Charlottenburg.
• A small team with a high degree of autonomy and quick decision-making processes.
• Agentic coding with tools like Claude and Gemini is part of our everyday development routine.
• Good English is sufficient to work with us. Knowledge of German is welcome but not required.
• An MDR-compliant environment where the systems we develop are actually used in hospitals and must function reliably.

Sound interesting? Then send us your resume and your salary expectations to jobs@asclea.ai. We don’t need a cover letter - we’ll discuss everything else directly.