Machine Learning Engineer, Model Productionization | Swiss RegTech | Zurich (Hybrid)

Machine Learning Engineer, Model Productionization | Swiss RegTech | Zurich (Hybrid)

A Swiss company with nearly four decades of engineering pedigree behind the software that fights financial crime inside 1,500+ banks across 80 countries is looking for the engineer who sits between its data science team and production. With the data scientists building new cutting edge, real-time models, what's needed is the person who takes those and makes them run, fast, reliably, and at scale, inside a Kubernetes-native production environment built for real-time decisioning.

This is not a modeling role or a data science hire. It's an engineer with real production software discipline who's comfortable being the one who makes someone else's model actually work under load, can pull together the resources needed, and be the glue.

Responsibilities, updated:

• Take models built by the data science team from prototype to live production, owning the engineering that gets them there.
• Build and optimise real-time inference pipelines where latency and throughput are the constraint, not just accuracy.
• Own Spark-based data engineering pipelines feeding model training and inference.
• Deploy, run, and troubleshoot workloads on Kubernetes in production.
• Translate model and product direction into a concrete technical plan for the dedicated engineering resources, and lead in execution day to day.
• Work directly with data scientists, translating experimental or research-grade code into tested, maintainable, production software.
• Contribute to the CI/CD and monitoring practices around model deployment.

Requirements, updated:

• A software engineering background first where you have shipped production systems
• Hands-on Spark experience at real pipeline scale, not project or PoC work.
• Production Kubernetes experience.
• Direct experience taking ML models from a data science handoff into live production, with real exposure to the performance and latency trade-offs that involves.
• Comfortable taking ambiguous or incomplete direction from stakeholders and turning it into a clear technical plan and task breakdown for a small dedicated team, rather than waiting for fully-specified requirements.

Nice to have, unchanged:

• Kafka production experience.
• Exposure to fintech, financial services, or another regulated industry.
• Familiarity with model lifecycle tooling (MLFlow or equivalent) or serving frameworks.