MLOps Engineer

Role Description

Build and operate the self-service AI/ML platform of the European Data Platform on AWS, so that platform users can deploy models, functions and GenAI agents by themselves, securely and through standard CI/CD pipelines.

Key Objectives 

• Build and maintain the platform as a product: templates and CI/CD pipelines that let users release without tickets.
• Define all infrastructure as code and deploy it through pipelines across accounts and environments.
• Apply security by design: least-privilege access, secrets management and policy as code.
• Operationalize ML models and enable GenAI use cases (agents, tools, knowledge bases) on the platform.
• Monitor the platform, control costs and keep the documentation up to date for users.

Main Responsibilities  

• Work with the security, networking and platform teams to guarantee a secure and stable environment.
• Work with the Data Architect to keep development aligned with the target architecture.
• Work with other MLOps developers and platform users to take ML and GenAI solutions to production.

This is you

• You are enthusiastic about working in a European team on a common Data Platform.
• You think about the people who use what you build and use AI tools in your daily work.
• You are organized and get things done

Must have:

• Platform / DevOps engineering experience: building tools, templates or pipelines used by other teams (self-service, golden paths).
• AWS in a multi-account setup: IAM (roles, resource policies, cross-account access), VPC networking, KMS, Secrets Manager, S3, Lambda, API Gateway, Step Functions, ECR, ECS/Fargate and CloudWatch.
• Infrastructure as code: AWS CDK (Python preferred) or CloudFormation. Terraform experience is also valued.
• CI/CD: GitLab CI (or equivalent), Git and Gitflow, with deployment by environment.
• Python with software engineering practices (tests, packaging, debugging) and Docker.
• Good level of English to work with teams from other countries.
• Experience working with AI coding assistants such as Claude Code or GitHub Copilot.

Nice to have:

• MLOps: SageMaker (Pipelines, Model Registry, endpoints, Feature Store), MLflow.
• GenAI on AWS: Bedrock, AgentCore, MCP, RAG / Knowledge Bases, agent evaluation.
• Policy as code and service-to-service authentication and authorization (OAuth, Cognito).
• Observability and troubleshooting of distributed systems (network, DNS, TLS, IAM).
• Kubernetes / EKS.
• Data tools: Athena, Glue, Spark / PySpark, Collibra.