MTS - AI / MLOps

Everpure, formerly Pure Storage, is hiring a Member of Technical Staff for AI and MLOps into its Pure Solutions group in Bangalore. The team connects Everpure storage platforms to the wider machine learning tooling ecosystem so that enterprise AI workloads run dependably in production. You would own the reference architectures and automation that show how the company's data plane supports training and inference at scale.

What you would work on - Building automated machine learning pipelines that span orchestration platforms and CI/CD tooling, with Everpure arrays acting as the underlying data layer - Producing validated architectures for distributed training clusters, defined entirely through infrastructure as code - Tuning GPU based model serving for throughput and low latency, including inference server configuration - Standing up lab environments and writing the supporting technical material that sales teams and partners rely on - Working alongside data scientists and product management to shape where the AI platform goes next - Investigating how high speed networking and fast shared storage affect end to end training performance

What they are looking for - Practical background designing MLOps pipelines and managing infrastructure declaratively - Expert Python skills applied to data science and machine learning operations work - Depth in PyTorch, including distributed training spread across multiple GPUs - Strong Kubernetes knowledge, particularly scheduling and allocating GPU resources - Familiarity with CUDA, GPU serving stacks, high performance storage and data centre infrastructure - Willingness to work from the Bangalore office, since the posting describes this as primarily an in office role

Stack: Kubeflow, MLflow, Ray, FlashBlade, FlashArray, Portworx, Git, Jenkins, Ansible, Terraform, Python, pandas, NumPy, PyTorch, NVIDIA Triton Inference Server, DeepSpeed, Megatron-LM, Vertex AI, SageMaker, CUDA, Kubernetes, RoCE, 100GbE