Deskripsi Pekerjaan
Apex Neural Dynamics is seeking a Senior AI Infrastructure Engineer to design, scale, and optimize our distributed GPU training clusters and inference pipelines. In this role, you will eliminate bottlenecks at the intersection of hardware, network, and distributed systems software to accelerate production-grade foundation models.
Tanggung Jawab
- Architect and maintain high-throughput distributed training infrastructure using Kubernetes and bare-metal nodes.
- Optimize GPU utilization, networking topologies, and storage throughput for multi-node model training.
- Implement automated telemetry, fault detection, and self-healing systems across large-scale compute fleets.
- Collaborate with ML research teams to debug distributed training bottlenecks and profile CUDA kernels.
- Establish robust CI/CD and deployment frameworks for ultra-low-latency model inference APIs.
- Enforce infrastructure-as-code practices, zero-trust network security, and fine-grained access policies.
Kualifikasi
- 5+ years of experience engineering Linux-based distributed infrastructure and Kubernetes in production.
- 3+ years hands-on experience provisioning and tuning NVIDIA GPU clusters (A100/H100) and CUDA workloads.
- Proficiency in Go, Python, or Rust for systems software and operational tooling.
- Deep working knowledge of high-performance networking fabrics including InfiniBand and RoCE.
- Experience configuring and benchmarking distributed file systems such as Ceph, Lustre, or GPFS.
- Demonstrated capability in debugging complex kernel-level, hardware, or network performance anomalies.