SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Salary: USD 166,000 - 220,000 / annual
Anduril Industries is a defense technology company building advanced military systems powered by Lattice OS, an AI-powered operating system for real-time command and control. The CorpTech Infrastructure Engineering team builds and operates foundational infrastructure enabling engineers, researchers, and product teams across the company to deploy fast, scalable systems.
You will lead the vision, execution, and long-term stability of Anduril's GPU training infrastructure at scale. This is a highly hands-on role focused on cluster robustness, ensuring high-performance GPU systems are highly available, fault-tolerant, and resilient for ML platform and research teams company-wide. Your primary focus is logical stability and automated resilience—building self-healing mechanisms to proactively detect and isolate hardware faults, tuning NCCL and high-speed networking, and optimizing Kubernetes, Run:AI, and Ray scheduling. By replacing manual triage with automated deployment tooling and deep observability, you will ensure massive-scale training infrastructure runs seamlessly and scales without linear headcount growth.
Key responsibilities include:
- Rack, stack, cable, and bring up GPU compute (H200/B200/B300, NVL72) including physical topology, power, cooling, firmware/BIOS, and burn-in validation
- Build and tune the interconnect fabric (NVLink, InfiniBand, RoCE, Spectrum-X) connecting hundreds of GPUs into low-latency training and inference clusters
- Integrate high-performance parallel storage (VAST, DDN, Weka) to sustain throughput demanded by distributed training and terabyte-scale multi-modal datasets
- Automate cluster deployment and configuration end-to-end, including infrastructure as code for bring-up, firmware/driver management, and fabric configuration
- Operate and extend Kubernetes/Run:AI environment for GPU scheduling, quota management, and multi-tenant workload isolation
- Own fleet health: monitoring, alerting, and rapid triage of hardware and network faults (bad transceivers, GPU Xid errors, NCCL/collective failures, RoCE congestion)
- Onboard engineers and researchers onto the platform and act as escalation point for debugging, training, and optimizing workloads
- Partner with product-facing teams to understand emerging compute needs and translate them into platform capability
REQUIREMENTS:
- 10+ years in hands-on infrastructure, HPC, or datacenter engineering role supporting GPU compute at scale
- Hands-on experience with H200/B200/B300 (or comparable) GPU systems: bring-up, cabling, firmware/driver management
- Experience with high-performance interconnects (NVLink, InfiniBand, RoCE, Spectrum-X) in clusters of hundreds of GPUs
- Experience with high-performance parallel storage (VAST, DDN, Weka, Lustre, or similar)
- Kubernetes required; Run:AI or similar GPU scheduling/orchestration experience strongly preferred
- Strong automation background; ability to build repeatable, automated deployment pipelines rather than manual processes
- Able to lift/move 50+ lbs and perform physical datacenter work (rack/stack/cable/troubleshoot)
- Eligible to obtain and maintain an active U.S. Top Secret clearance
PREFERRED QUALIFICATIONS:
- Experience with NVIDIA NVL72 rack-scale systems
- Experience supporting LLM token serving/inference infrastructure alongside training clusters
- Network fabric tuning experience (congestion control, adaptive routing, QoS) for RoCE/InfiniBand at scale
- Familiarity with GPU/network observability tooling (DCGM, fabric telemetry) and automated fault detection
- Experience supporting infrastructure as a shared platform serving multiple internal customer teams with differing requirements
