Embedded Autonomy Engineer

About the role

Mach Industries is a rapidly growing defense technology company focused on developing next-generation autonomous defense platforms. The Embedded Autonomy Engineer role involves bringing up Jetson-class compute, integrating high-bandwidth sensors, and optimizing the real-time sensor-to-GPU data path under real SWaP, thermal, and vibration constraints.

Key Responsibilities

• Own bring-up and lifecycle of the Linux mission computer: board support packages, device trees, kernel configuration, and bootloaders on NVIDIA Jetson/Tegra and similar SoCs.
• Integrate high-bandwidth sensors into Linux end to end: MIPI CSI cameras with GMSL/FPD-Link SerDes, V4L2, and ISP tuning, plus radar over PCIe or Ethernet; own time-sync and calibration across the sensor suite.
• Build and optimize the real-time sensor-to-GPU data path that feeds perception, localization, and inference: zero-copy, DMA, GPU memory management, and camera-to-inference latency.
• Stand up and maintain the on-target execution environment for autonomy and ML: CUDA/TensorRT runtime, GPU/CPU scheduling, edge containers, and the drivers and libraries the stack depends on.
• Define software-hardware interface specs with electrical engineers during board design cycles; debug integration with logic analyzers, oscilloscopes, and UART.
• Contribute to the autonomy runtime on the mission computer (middleware, health monitoring, state management) as the engineer who knows the Linux compute layer best.

Required Qualifications

• 5+ years of embedded Linux development on custom hardware, with deep expertise in device tree configuration, BSP customization, and kernel bring-up.
• Camera and sensor integration on Linux: MIPI CSI and V4L2, plus GMSL or FPD-Link SerDes; comfortable owning the sensor-to-memory path.
• Experience with Yocto or L4T-based build systems; fluency in C, C++, Bash, and Python.
• Strong hardware debugging with logic analyzers, oscilloscopes, CAN, and UART; reads schematics and collaborates fluently with EEs.
• Track record taking Linux-based compute platforms from bring-up to real-world deployment, with real-time and performance optimization.
• Wants to work up the stack into autonomy and on-target inference, not only maintain the OS and boards.

Preferred Qualifications

• NVIDIA Jetson and the L4T stack; CUDA, TensorRT, GStreamer/DeepStream for on-target inference and media pipelines.
• ROS 2/copper-rs/dora-rs and integration of autonomy, perception, or localization software on the mission computer.
• Multi-sensor time-sync and calibration (PTP, hardware triggering).
• GPU pipeline and memory optimization, DMA, and zero-copy under real-time constraints.
• Nix or NixOS workflows; strong opinions about Rust.
• Experience in contested or degraded environments (RF denial, low-light/night, high-vibration platforms).