Senior ML Platform Engineer (AI Infrastructure) | Swiss RegTech | Zurich
A Swiss RegTech company whose software fights financial crime inside 1,500+ banks across 80 countries is building out the platform layer behind its AI initiatives. Their AI Model Engineers build the models that power real-time fraud detection, AML scoring, and sanctions screening. What's missing is the person who builds the framework those models run on, the reusable inference and training infrastructure, the MLOps pipelines, the Kubernetes-native backbone that lets model work actually ship into production, including into air-gapped banking environments with or without GPU availability.
This is not a data science role and not a role for someone chasing model accuracy metrics. It is a platform and framework engineer with real software engineering discipline, someone who has built the plumbing that lets other people's models train, deploy, and run reliably at scale, inside infrastructure they do not always control.
What you'll do
- Design, build, and maintain the AI/ML infrastructure behind a microservice architecture running on Kubernetes.
- Build reusable AI inference and training framework components used by the model engineering team.
- Own the MLOps lifecycle: CI/CD pipelines for model training, testing, and deployment, using tools like MLFlow.
- Work across the backend stack: Kafka, gRPC, REST APIs, and data engineering pipelines built on Spark.
- Write and maintain production-grade Helm charts and DevOps routines, not just CLI-level Kubernetes usage.
- Support model deployment and performance in varied production environments, including isolated and air-gapped client infrastructure.
What you'll bring
- Financial crime domain knowledge (AML, fraud, sanctions screening, or KYC). This is a hard requirement, not a nice-to-have.
- 5+ years hands-on AI/ML platform engineering, with genuine depth in Kubernetes and Spark-based data engineering.
- Experience delivering AI/ML platform solutions into isolated, non-SaaS environments, ideally productionised software rather than internal tooling.
- Strong distributed systems fundamentals and clean, testable engineering practice.
- Familiarity with the broader AI/ML lifecycle: PyTorch, TensorFlow, ONNX, or Ray are all assets, no single one is a filter on its own.
- Experience in security-sensitive sectors (finance, health, defense, intelligence) is a strong plus beyond the core financial crime requirement.
Location: Zurich - 3 days a week on-site
Employment: Permanent full time
Language: English
Start: ASAP (however up to 3 month notice periods are also acceptable and understandable)