Deskripsi Pekerjaan
TensorScale Systems is seeking an exceptional Senior AI Infrastructure Engineer to architect, scale, and optimize distributed compute platforms powering state-of-the-art foundation models.
In this role, you will lead the orchestration of high-performance GPU clusters, design fault-tolerant distributed storage systems, and minimize training bottlenecks. You will work cross-functionally with ML research and platform engineering teams to maintain 99.99% infrastructure availability while continuously driving down training compute latency.
Tanggung Jawab
- Architect and maintain distributed compute platforms running on large-scale NVIDIA GPU clusters (H100/A100).
- Design, deploy, and optimize low-latency Kubernetes and Slurm scheduling pipelines for distributed training and inference workloads.
- Optimize high-throughput storage systems, network topologies (InfiniBand/RoCE), and collective communication libraries (NCCL).
- Implement comprehensive observability, telemetry, and automated remediation systems for large-scale GPU failures.
- Collaborate directly with ML researchers to profile training jobs, identify hardware bottlenecks, and accelerate iteration cycles.
- Establish infrastructure-as-code (IaC) pipelines using Terraform to support multi-region hybrid cloud deployments.
Kualifikasi
- 5+ years of software engineering experience focusing on systems, infrastructure, or cloud platforms.
- Demonstrated experience configuring and tuning large-scale GPU clusters for deep learning workloads.
- Deep operational expertise with Kubernetes, Linux kernel internals, and container runtimes at scale.
- Proficiency in Go, Python, or Rust, with strong systems programming fundamentals.
- Hands-on experience with InfiniBand, RoCE, or high-throughput network fabric engineering.
- Experience automating infrastructure via Terraform, Pulumi, or similar declarative tools.
- BS, MS, or equivalent experience in Computer Science, Computer Engineering, or related technical disciplines.