Lead Developer Infrastructure (DevInfra) Engineer, Robotics Software

📋 Description

• • Own and continuously evolve Bedrock’s GitLab CI/CD pipelines and automated test infrastructure, ensuring that developer velocity remains high as the AUV fleet scales from dozens to hundreds of autonomous nodes.
• • Design and maintain a pull-based deployment architecture that is resilient to intermittent satellite, LTE, and acoustic connectivity, enabling zero-touch rollouts to vehicles operating anywhere from the Arctic to the equator.
• • Collaborate daily with hardware, embedded, and autonomy teams to integrate simulation (Gazebo, Drake) and hardware-in-the-loop rigs into the release gate, cutting the number of costly sea trials per feature by at least 50 %.
• • Build deterministic build, test, and packaging workflows for heterogeneous ARM/x86 compute modules, CUDA-enabled GPUs, and micro-controller subsystems, using CMake, PlatformIO, Docker, and OCI artifacts.
• • Instrument every release with end-to-end observability—logs, metrics, distributed tracing—so regressions in navigation, perception, or power systems are detected within minutes, not days.
• • Create a feature-to-fleet tracking dashboard that maps Git commits to specific AUV serial numbers, firmware versions, sensor configurations, and mission history, giving product and operations teams a single source of truth for readiness.
• • Champion infrastructure-as-code practices using Ansible and Terraform, turning one-off snowflake testbeds into reproducible environments that any engineer can spin up in <15 minutes.
• • Drive cultural change by leading lunch-and-learns, brown-bags, and blameless post-mortems that raise the bar for automated testing, progressive rollouts, and rollback discipline.
• • Partner with ML and data teams to extend pipelines to on-vehicle model training and over-the-air model updates, ensuring that new perception models reach the fleet safely.
• • Contribute to open-source initiatives and publish post-mortems externally, positioning Bedrock as a thought leader in field robotics DevOps.
• • Within your first 30 days, ship a hot-fix to a live AUV via the new pull-based system; within 90 days, cut average release cycle time from weeks to days; within 6 months, make release day a non-event celebrated only by green dashboards.