SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
SumUp's Machine Learning Platform team builds foundational infrastructure that enables data scientists across the company to move models from idea to production faster. Currently, this process takes months rather than weeks, creating bottlenecks in fraud detection, lending decisions, and other critical models. You'll join a small, high-trust team of ML and MLOps engineers working alongside Data Streaming and Data Gateway teams, giving you visibility into how data moves and transforms across the business.
In this role, you will design and build ML platform tooling supporting feature engineering, training, experimentation, monitoring, and serving for both online and offline use cases. You'll simplify how data scientists create features by reducing reliance on complex Spark workflows through better abstractions and tooling. A key focus is standardizing ML infrastructure and developer experience, replacing fragmented ad hoc solutions with scalable, self-service components.
You'll partner closely with data scientists to understand their workflows and pain points, translating these into concrete platform improvements. You'll mentor data scientists on best practices and help drive adoption of the platform across teams. Collaboration with Data Platform teams will ensure ML initiatives remain aligned with wider data engineering work.
The team sits within SumUp's broader fintech platform, which serves over 4 million businesses across 39 markets. This is an office-first role based in Berlin.
REQUIREMENTS:
- More than 6 years of experience building production-grade ML infrastructure such as feature stores, training or orchestration frameworks, experimentation platforms, or model serving and monitoring systems
- Strong proficiency in Python and ML libraries
- Experience with ML orchestration tools such as Metaflow, and familiarity with platform tooling such as MLflow, Kubeflow, Tecton, Chronon, or Airflow
- Solid understanding of batch and real-time data processing patterns (e.g., Spark, Flink, Kafka)
- Comfort working with AWS cloud-native services and infrastructure tooling such as Docker, Kubernetes, or Terraform
- Track record of partnering with stakeholders to turn workflow pain points into reusable, self-service platform capabilities
