Senior Machine Learning & AI Engineer (m/f/d)

Do you want to redesign how a company runs with AI, from how we find customers to how we manage residual values and answer service requests? Do you enjoy moving between classic ML and LLM agents, and want to shape an AI team in an environment where your decisions stick?

We are looking for a Senior Machine Learning & AI Engineer (m/f/d) to build ML and agent solutions on AWS and Snowflake and to set the technical foundation of a growing AI team.

Bike Mobility Services GmbH, a subsidiary of the Dutch company Pon.Bike, is part of a group that owns international brands like Kalkhoff, Cervélo, Urban Arrow, Cannondale, Santa Cruz, Focus, and Royal Dutch Gazelle. Our mission is to get as many employees as possible onto bikes for their commute, through our digital Lease a Bike portal. Data already drives how we steer our portfolio, manage residual values, and find new customers.

You will take a business process, understand where the potential lies, and rebuild it with AI: predictive models where decisions need numbers, agents where people spend their day on repetitive work.

• Predictive ML: Build models for lead generation, demand forecasting, residual value forecasting, portfolio steering, and fraud detection.
• Agents & Generative AI: Redesign workflows in customer service and other teams with RAG and agent-based solutions, integrated securely into our systems and evaluated for output quality.
• MLOps / LLMOps: Build and run ML pipelines on AWS and Snowflake. Monitor performance, drift, and cost, and decide when a model needs retraining or should be retired.
• Business Partnership: Turn business needs into measurable use cases together with stakeholders and the data engineering team, and own delivery end to end.
• Team Building: Evaluate new models and frameworks, shape shared standards and tooling, and mentor more junior colleagues as the team grows.

Across all of this, you keep responsible AI in view: data privacy, bias, and explainability.

Must-Haves:

• Experience: 5+ years in Data Science, Machine Learning, or AI Engineering, with at least one solution you took into production and ran there. You are strong in either predictive ML or LLM applications, with hands-on experience in the other.
• Python, SQL & ML Fundamentals: Very good Python and SQL, ML frameworks like scikit-learn or PyTorch, and sound experiment design and model evaluation.
• Production on AWS: Deploying and operating ML solutions on AWS (e.g. SageMaker, MLflow, Docker, CI/CD), with Snowflake or a comparable data platform.
• Generative AI & Agents: Practical experience with RAG, tool calling, and agentic workflows, including how to evaluate LLM output quality, and daily use of AI coding tools like Claude Code or GitHub Copilot.
• Ownership: You make design decisions you can defend, guide other engineers, explain trade-offs to non-technical colleagues, and are fluent in English.

Nice-to-Haves:

• Process Redesign: Experience rebuilding a business process with AI together with the teams who run it.
• Leasing Domain: Experience with leasing, subscription, or other asset-based businesses.
• Languages: Conversational German or Dutch.

• Business Impact: Your work ends up in processes people use every day, with direct access to the teams who run them.
• Early-Stage AI Team: Shape how the team builds, what it uses, and which problems it takes on first.
• AI-Native Engineering: AI coding tools and agents in your daily workflow, and time to test new models and frameworks against real use cases.
• Bike Leasing & Home Office: We subsidize your bike lease and bike purchase, and we equip your home office so you can work comfortably.
• Learning: You get a learning budget and conference days to grow where you want.
• 100% Remote: You work from anywhere in the EU, and we meet regularly as a team in one of our operating countries, so you get to know the people behind the use cases.

Send your CV and a few lines about a project you took into production to carina.wegmann@pon.com. The process has two steps: an introduction interview, then a technical conversation. We aim to give you feedback within two weeks of each step.