Staff AI Infrastructure Engineer

Responsibilities

• Own the technical roadmap and architecture for Anduril’s foundational ML training, orchestration, experimentation, and MLOps platform.
• Build automated tooling for hyperparameter tuning, model profiling, experimentation tracking, and research-lifecycle optimization.
• Design and scale ETL pipelines processing terabytes of multimodal sensor and simulation data, including video, camera feeds, radar, flight telemetry, and simulation logs.
• Architect high-throughput, low-latency model-serving infrastructure for cloud and air-gapped tactical edge environments.
• Build ML CI/CD pipelines with automated validation, canary deployments, and rollback capabilities.
• Implement continuous model evaluation, validation, and reinforcement learning alignment loops using RLHF and DPO.
• Set unified infrastructure standards with AI researchers, computer vision teams, and platform engineers.
• Lead technical direction and complex migrations, mentor engineers, and help recruit and expand the infrastructure team.

Requirements

• 7+ years of software engineering experience designing, building, and operating production-scale machine learning systems and MLOps platforms.
• Proficiency in Python, Go, C++, or similar backend languages, with deep knowledge of ML systems design, memory management, and distributed computing.
• Deep experience with Docker, Kubernetes, GPU scheduling and orchestration, and distributed training frameworks such as PyTorch Distributed, Ray, Slurm, or Megatron-LM.
• Hands-on experience building distributed ETL pipelines and managing terabytes of unstructured or multimodal sensor data.
• Experience setting technical direction, leading complex system migrations, and mentoring senior engineers.
• Eligibility to obtain and maintain an active U.S. Top Secret security clearance.
• Preferred experience with secure, air-gapped, highly regulated, IL5/IL6, or GovCloud environments.
• Preferred experience building LLM, generative AI, or reinforcement learning training and evaluation platforms.
• Preferred experience profiling training hardware and optimizing network, memory, and accelerator utilization.
• Preferred experience designing multi-tenant ML platforms with resource isolation, quota management, and fair scheduling across shared GPU clusters.
• Preferred experience using custom AI accelerators such as AWS Trainium, Google TPUs, or custom ASICs.
• Preferred experience building ML observability systems with prediction drift detection, data quality monitoring, and automated retraining triggers.

Benefits

• Comprehensive competitive benefits are available at little to no cost to full-time employees.
• Full-time employees receive health and recovery support and other benefits.
• The role includes competitive equity grants in the majority of full-time offers.