MLOps Engineer - GPU Platform (Inference & Training)

Tasks

• Automate deployments with GitOps using ArgoCD and Helm
• Collaborate on training runs and model serving rollouts
• Implement GPU aware scheduling and warm pools
• Improve GPU utilization and reduce idle allocations
• Manage Talos Linux and Sidero Omni cluster lifecycle
• Monitor GPU health and resolve node incidents
• Operate NVIDIA GPU drivers and GPU Operator
• Optimize distributed training clusters
• Provision infrastructure with Terraform
• Scale inference with KEDA and Kafka autoscaling
• Set observability with Prometheus and VictoriaMetrics
• Track SLOs and GPU cost per generation
• Tune NCCL and GPU network performance

Perks/Benefits

• On site office relocation

Skills/Tech-stack

Amazon EKS | ArgoCD | Bottlerocket | CUDA | DCGM | GPU Operator | GVisor | GitOps | Go | Grafana | Helm | Infiniband | Istio | KEDA | Kafka | Karpenter | Kubernetes | Linux | Linux Networking | NCCL | NVIDIA GPU | NVIDIA GPU Operator | Nebius | Prometheus | Python | RoCE | Sidero Omni | Talos Linux | Terraform | Terraform Cloud | VictoriaMetrics