Skip to main contentEnglishEnglishBack to all jobsSenior Data Scientist (m/f/d)Germany, Munich, Germany, BerlinFull-timePermanent employeeApply for this jobYour daily workoutPut your data science and machine learning talent behind a good cause and help make the world a healthier place by developing the next generation of product innovations at EGYM. You will take a leading individual-contributor role on EGYM Genius – the engine that powers personalized training plans for our users – combining LLMs, classical machine learning, and agentic systems from idea all the way to production. Apply now as a Senior Data Scientist (m/f/d) and be part of our Data Science team in Munich or Berlin.Model Development: You use EGYM’s diverse training and health data to build machine-learning models that support effective and healthy workoutsData Exploration: You analyze complex datasets, validate hypotheses, and develop new approaches for challenging problemsLLM & Agent Design: You design and maintain LLM-based components and agents that reason over user context, call tools/APIs, and orchestrate multi-step flows to generate, adapt, and explain training plansLLM Evaluation: You set up robust evaluation pipelines for LLM features and agents (golden datasets, automatic metrics, human evals, regression checks, and safety/guardrail tests)Local & Open-Source Models: You are comfortable going beyond “just an API” - you can work with local/open-source models (running them on our infrastructure, adapting or fine-tuning them, and balancing quality, latency, and cost)Collaboration: You work closely with Software Engineering and Product Management while sharing feasibility insights and ongoing resultsImplementation: You develop solutions in a dynamic software environment aligned with the EGYM Tech RadarInnovation Monitoring: You track advancements in AI and machine learning to ensure our solutions remain state-of-the-artYour fitness levelExperience: You bring 5+ years of experience delivering machine-learning-based solutions, ideally in personalization, recommender systems, or decision-support productsProgramming & Data: You write production-ready Python and work confidently with structured and unstructured data across SQL and NoSQL databases; you’re comfortable collaborating with backend engineers working in JavaML & LLMs: You have strong ML fundamentals and hands-on experience building and evaluating LLM-powered features or agents (prompt design, retrieval, tool use, multi-step flows, defining quality criteria, building test sets, running automated and human evaluations, and doing error analysis). You’re also familiar with adapting local/open-source models and understand trade-offs in size, serving, performance, and costHybrid Approaches: You’re comfortable combining classical ML and LLMs (e.g. for scoring, ranking, recommendation, or as components in larger decision systems)Communication & Ownership: You explain complex topics clearly, work effectively with different stakeholders, and see problems through from definition to production launchYour training goal for your first 6 monthsDomain Knowledge: You become highly familiar with EGYM’s health and training dataCross-Functional Experience: You complete a rotation in a software development team and collaborate directly with engineersImpact Contribution: You deliver visible product improvements and optimize your work using real usage feedbackProcess Optimization: You help us to standardize how we build and ship models and LLM agents – defining reusable evaluation suites, automating training and deployment steps, and setting up monitoring for quality, drift, and safetyProduct Partnership: You act as a sparring partner for Product Management and help extend our product visionThe equipment we provide Modern Culture: Be part of a modern and international company culture where talent and passion are welcomed, heard, and part of the decision-making processContinuous Development: Develop your skills with varied,...
Want jobs like this matched to you?
SimpleCareer scores fresh postings against your résumé so you only see the matches that matter.