Overview
Dev.Pro Bogota, D.C., Capital District, Colombia — We invite a skilled Kubernetes Developer to join our fully remote, international team.
In this role, you'll build and optimize the Kubernetes orchestration platform and develop custom operators to run HPC/AI workloads efficiently on GPU clusters.
You'll enhance infrastructure performance and reliability, create internal tools to improve the developer experience, and ensure multi-tenant HPC workloads remain secure and compliant.
What’s in it for you
- Work on cutting-edge GPU infrastructure and next-gen HPC/AI workloads
- Build a Slurm-on-Kubernetes product from scratch and shape its architecture
- Collaborate with a top-tier international team and grow through continuous learning and conference participation
Key Responsibilities
- Design, develop, and manage Kubernetes platforms for GPU-intensive AI/HPC workloads
- Design and build a Slurm-like orchestration layer on Kubernetes for HPC/AI workloads
- Develop custom operators and controllers for GPU job scheduling and execution
- Integrate batch schedulers with Kubernetes to provide a hybrid HPC/Cloud product
- Implement advanced GPU resource management and multi-tenant isolation policies
- Build internal tools and a self-service platform to simplify AI/HPC job deployment and management
- Monitor GPU clusters, troubleshoot production issues, and ensure high availability, fault tolerance, and disaster recovery
- Develop CI/CD pipelines for GPU-intensive workloads
- Ensure compliance with data sovereignty and international regulations
Qualifications
- 3+ years of hands-on Kubernetes experience in production
- Experience with HPC schedulers (Slurm, PBS, LSF, Volcano)
- Strong background in GPU resource management and distributed systems
- Experience with cloud/hybrid cloud architectures (AWS, GCP, Azure, on-prem GPU clusters)
- Knowledge of Kubernetes operators, CRDs, scheduling, networking, and storage
- Deep knowledge of HPC job scheduling and workload orchestration
- Expertise in IaC (Terraform, Helm, or GitOps: ArgoCD/Flux) and monitoring & observability (Prometheus, Grafana, Jaeger, ELK)
- Programming skills in Go, Python, Bash/Shell
- Familiarity with PyTorch, TensorFlow, distributed training, and model serving
- Skills in Linux administration, performance tuning, and advanced networking (RDMA, InfiniBand, TCP/IP, DNS, load balancing)
- Experience in storage management and optimization for large datasets
Note: This role is fully remote and international, with a focus on collaboration across time zones.
#J-18808-Ljbffr