28 days ago
London, United KingdomSenior
Responsibilities
- Provision and manage cloud-native AI/ML infrastructure using Kubernetes, Docker, and GPU orchestration frameworks such as NVIDIA GPU Operator, Slurm, or Ray.
- Automate platform infrastructure with Terraform, Helm, and Ansible.
- Optimize GPU compute workloads, high-speed networking, and storage for model training and low-latency inference.
- Build and maintain CI/CD and MLOps pipelines for model training, evaluation, packaging, and production deployment.
- Deploy large language models and generative AI workloads using Triton Inference Server, vLLM, or TensorRT-LLM.
- Automate model validation and monitor model drift, data drift, and latency bottlenecks.
- Monitor and optimize cloud spending across AWS, GCP, and Azure GPU and CPU clusters.
- Implement auto-scaling, spot instance policies, and dynamic resource allocation.
- Establish benchmarking and telemetry for AI model training and serving unit economics and throughput.
- Implement end-to-end observability with Prometheus, Grafana, OpenTelemetry, Weights & Biases, or MLflow.
Requirements
- Production experience in DevOps, Site Reliability Engineering, or Platform Engineering, including experience with AI/ML infrastructure.
- Experience deploying, scaling, and operationalizing machine learning models and LLMs in cloud-native production environments.
- Experience managing compute-intensive GPU infrastructure and high-performance computing environments.
- Advanced proficiency with Kubernetes, Docker, Helm, KubeFlow, and service meshes such as Istio.
- Hands-on experience with Terraform, Ansible, GitHub Actions, ArgoCD, or Jenkins.
- Experience with vLLM, Ray, MLflow, LangChain, LangSmith, DeepSpeed, or Hugging Face TGI.
- Experience with AWS, GCP, Azure, Kubecost, and GPU cost optimization techniques.
- Strong skills in Python, Bash, or Go, with deep knowledge of Linux kernel tuning and performance monitoring.
Benefits
- Inclusive workplace focused on diversity, equity, inclusion and belonging.
- Reasonable accommodations are available for applicants and employees with disabilities.
Tech Stack
AnsibleAWSAzureBashDockerGitHub ActionsGoGoogle Cloud PlatformGrafanaHelmIstioJenkinsKubernetesLinuxMLflowPrometheusPythonTerraform
Categories
About Anaplan
Anaplan builds a cloud platform for enterprise planning, forecasting, and scenario modeling used by finance, supply chain, sales, and workforce teams. It sells subscription software and related services to large organizations that need connected, cross-functional planning. Founded in 2006 and headquartered in Miami, Florida, Anaplan was acquired by Thoma Bravo in 2022 after previously trading on the NYSE.