Senior AI Storage Architect for GPU-Accelerated Training

Bitdeer Technologies Group; San Jose (CA); On-site; USD 180,000 - 240,000; Full time

Job summary

Bitdeer seeks a Senior AI Storage Infrastructure Engineer to build the data-delivery fabric for AI-native NeoCloud. You will architect high-performance storage, enable GPUs with data, and design NVMe caching for large model weights in distributed training and inference environments.

Role sits at the intersection of distributed storage, kernel-level I/O, and Kubernetes orchestration, requiring strong design and collaboration across storage, GPU, and network teams.

Qualifications

• 5+ years of experience in distributed storage systems and high-performance file systems.
• Deep knowledge of POSIX I/O semantics and file systems.
• Strong experience with Kubernetes CSI and storage operators.
• Linux kernel I/O and performance tuning expertise.
• Familiarity with RDMA, InfiniBand, and RoCE and related storage subsystems.

Responsibilities

• Design, deploy, and maintain high-performance CSI drivers for parallel file systems.
• Architect GPUDirect Storage integrations to enable DMA between NVMe drives and GPU memory.
• Develop local NVMe caching strategies for large model weights and datasets.
• Optimize IOPS, throughput, and latency across the containerized storage stack.
• Collaborate with GPU Systems & Fabric teams to optimize for RDMA and InfiniBand interconnects.
• Implement automated monitoring and alerting for storage performance and degradation.

Skills

Distributed storage

Kubernetes CSI

Linux kernel I/O

RDMA networking

Storage performance tuning

CI/CD pipelines

Education

Bachelor’s or Master’s degree in Computer Science or Electrical Engineering

Tools

Terraform

Ansible

Job description

Bitdeer seeks a Senior AI Storage Infrastructure Engineer to build the data-delivery fabric for AI-native NeoCloud. You will architect high-performance storage, enable GPUs with data, and design NVMe caching for large model weights in distributed training and inference environments.

Role sits at the intersection of distributed storage, kernel-level I/O, and Kubernetes orchestration, requiring strong design and collaboration across storage, GPU, and network teams.