Deskripsi Pekerjaan
Nexus Compute Labs is looking for a Senior AI Infrastructure Engineer to design, scale, and optimize our distributed GPU training and inference clusters. You will build the foundational systems supporting multi-billion-parameter foundation models, working alongside top researchers to maximize throughput and eliminate hardware bottlenecks.
This role offers high visibility, cutting-edge scale, and the opportunity to shape critical AI delivery infrastructure from day one.
Tanggung Jawab
- Architect and maintain bare-metal and cloud-based multi-node GPU clusters (NVIDIA H100/A100).
- Design automated provisioning, orchestration, and monitoring tooling using Kubernetes, Slurm, and Terraform.
- Optimize high-throughput network architectures leveraging InfiniBand, RoCE, and NCCL.
- Implement observability and automated alerting pipelines for cluster health, job preemption, and silent hardware failures.
- Collaborate with ML research teams to debug distributed training bottlenecks across PyTorch and DeepSpeed workloads.
- Lead disaster recovery, capacity forecasting, and security compliance across all compute platforms.
Kualifikasi
- 5+ years of experience managing production-scale Linux infrastructure and container orchestration.
- 3+ years operating large-scale distributed training clusters with Slurm or Kubernetes.
- Expert proficiency in Python, Go, or Rust for infrastructure tooling and automation.
- Deep understanding of high-speed interconnects (InfiniBand, RoCE) and GPU networking primitives (NCCL).
- Demonstrated experience with Infrastructure-as-Code (IaC) using Terraform, Ansible, or Pulumi.
- Strong grasp of distributed storage solutions including Lustre, Ceph, or GPFS.