Platform Engineer

CuspAIAmsterdamOn-siteFull-time2 weeks ago

CuspAI is seeking a Platform Engineer with strong Python coding, DevOps and cloud platform expertise to build and maintain the ML infrastructure platform that powers their AI research teams. The role involves designing cloud-native Kubernetes platforms, owning the MLOps stack, scaling distributed ML training, and ensuring platform reliability. The ideal candidate has deep experience with ML workloads on Kubernetes, multi-cloud platforms, Infrastructure as Code, and GPU infrastructure.

What you'll do

Build the ML Infrastructure Platform: Design and implement a cloud-native platform on Kubernetes that enables researchers to easily train, evaluate, and deploy models without worrying about infrastructure complexity.
Own the MLOps Stack: Implement and maintain CI/CD pipelines, model registries, experiment tracking systems, and deployment automation - the full lifecycle of ML operations.
Scale Distributed ML Model Training: Build infrastructure to support distributed training across many GPUs, implementing solutions for data pipelines, checkpointing, and resource optimisation.
Platform Reliability: Ensure 99.9% uptime for our ML platform through monitoring, alerting, and automated recovery systems. Be the guardian of infrastructure stability.
Configuration Management: Use Kapitan to manage complex multi-environment Kubernetes configurations and ensure consistent deployments across dev, staging, and production.
Cost Optimisation: Implement resource management strategies to optimise cloud spending and detect inefficient usage while maximising computational throughput for our research teams.
Developer Experience: Create tools and abstractions that make it simple for researchers to go from experiment to production without deep infrastructure knowledge.
Interdisciplinary Collaboration: Partner closely with ML Researchers, Chemists, Materials Scientists and Software Engineers to understand their needs, and to build the infrastructure for groundbreaking projects

What you need

Must-have

PythonGoKubernetesAWSGCPTerraformHelmKapitan

Nice-to-have

FlytePantsSlurmPBSTritonTorchServeKServeHPC environments

Soft skills

collaborationautonomycommunicationproblem solvingmentoringadaptability

Platform EngineerML InfrastructureDevOpsMLOpsKubernetesAIMaterials ScienceCloudPythonGo

/What you'll do

/What you need

Must-have

Nice-to-have

Soft skills

What you'll do

What you need