AI Infrastructure Engineer

The future of defense will be decided by those who field intelligent machines at scale. At Scout, we’re developing Fury — the first robotic foundation model for defense — to give U.S. forces overwhelming, adaptable, and autonomous power across every domain. Fury enables human operators to command fleets of robots through natural language, and empowers those machines to sense, decide, and act together as one. It’s not just a leap in autonomy, it’s a force multiplier built for real-world conflict. This mission will ask everything of us: urgency, precision, and relentless work.The RoleWe’re looking for an AI Infrastructure Engineer to build and scale the backbone of Fury’s model training and deployment ecosystem. You’ll design the data, compute, and orchestration infrastructure that enables our vision-language-action models to learn from massive real-world datasets and operate across edge and cloud environments. This role bridges systems engineering, distributed computing, and machine learning infrastructure. Your work will ensure our teams can iterate rapidly, train large models efficiently, and deploy them reliably on robotic platforms in the field.We’re a startup. You’ll be moving fast, context-switching daily, and helping define the culture and process as we go. This is a rare opportunity to come in early and architect the future of defense.ResponsibilitiesDesign and implement data pipelines for ingesting, transforming, and storing petabytes of multimodal data from Fury’s robotic and operator systemsDevelop internal tooling for dataset exploration, curation, versioning, and quality monitoring over timeBuild and maintain distributed training infrastructure (cloud and on-prem) for large-scale multimodal and foundation model trainingImplement job orchestration workflows for launching, tracking, and debugging large-scale model runsIdentify and remediate bottlenecks in compute, memory, storage, and network performance to optimize throughput and cost efficiencyCollaborate with AI, autonomy, and systems teams to ensure data and training infrastructure supports real-time and mission-critical use casesMaintain observability and reliability tooling for training and inference pipelinesStay current on best practices in MLOps, distributed training frameworks, and AI infrastructure at scaleQualifications3+ years of experience in ML infrastructure, MLOps, or large-scale data systemsProven experience with distributed training (PyTorch DDP, DeepSpeed, Ray, or similar) and workflow orchestration (Kubernetes, Airflow, or equivalent)Strong proficiency in Python and cloud-native infrastructure (AWS, GCP, or Azure)Deep understanding of data engineering (ETL pipelines, object storage, data versioning, metadata management)Familiarity with containerization and deployment (Docker, Kubernetes) and monitoring systems (Prometheus, Grafana)Experience optimizing GPU cluster utilization, scaling training jobs, and profiling model performanceBachelor’s degree or higher in Computer Science, Electrical Engineering, or related technical fieldBonus: Experience with edge-deployed ML systems, federated training, or robotic data collection pipelinesMust be a U.S. Person due to required access to U.S. export controlled information or facilitiesWhy Join ScoutWork on the world’s most important frontier, ensuring U.S. and allied dominance in the age of intelligent machinesBe a core part of a team building the first defense-specific robotic foundation modelCollaborate with some of the top engineers in autonomy, AI, and national securitySee your work deployed on real systemsHelp define the future of intelligent defense systemsBacked by Draper Associates, Booz Allen Ventures, and other top investorsBenefitsCompetitive base salary and meaningful equityPremium medical, dental, and vision plans with $0 paycheck contributionCompetitive PTO and company holiday calendarCatered lunch daily and fully stocked kitchenEV chargingRelocation assistance (depending on role eligibility)