Embedding System Engineer
Job Description: Embedded System Engineer (BSP & Edge AI)
Position: Embedded System Engineer (BSP)
Department: R&D / Hardware Engineering
Location: On-Site / Hybrid (Part Time/Full Time)
Experience: 0-5 years (colleges students and freshers are encouraged to apply)
Degree: Preferred Masters but not required
About the Role
We are building OneBit AI, a sovereign, air-gapped AI computing appliance for defense and critical infrastructure. We are looking for an Embedded Linux Engineer to "build the brain" of our hardware. You will be responsible for the Board Support Package (BSP), kernel customization, and the low-level OS layer that allows our high-performance AI models to communicate with the hardware (NPU) and the host user.
You will turn a raw development board (RK3588) into a polished, plug-and-play USB AI accelerator.
Key Responsibilities
Board Support Package (BSP) Management:
Bring up and customize Linux operating systems (Ubuntu/Debian/Buildroot) for ARM-based SoCs (specifically Rockchip RK3588).
Manage bootloaders (U-Boot) and partition tables for reliable booting.
Kernel & Driver Development:
Configure and compile the Linux Kernel to enable specific hardware features.
Critical Task: Implement and debug USB Gadget Mode (g_ether / RNDIS) to allow the device to act as a virtual network interface over USB-C.
Optimize Device Tree overlays (.dts) to enable/disable peripherals for power saving.
System-Level Automation:
Write and maintain systemd service units to ensure mission-critical AI processes (LLM engines, API servers) auto-start under 45 seconds.
Implement "Watchdog" scripts to auto-restart services if they crash during heavy inference.
Hardware Acceleration Integration:
Integrate vendor-specific NPU drivers (rknn, librknnrt) into the OS.
Ensure the NPU memory allocation (cma) is optimized for large model inference.
Technical Skills (Must Have)
Core OS: Strong proficiency in Linux Systems Programming (C/C++, Bash, Python).
Hardware: Hands-on experience with ARM64 SoCs (Rockchip, NVIDIA Jetson, or Raspberry Pi CM4).
Kernel: Experience with configfs, libcomposite, and configuring USB Gadget subsystems.
System Management: Mastery of systemd, journalctl, and init processes.
Networking: Understanding of TCP/IP, static IP configuration, and bridging network interfaces in Linux.
Preferred Skills (Bonus)
Experience with NPU Toolkits (Rockchip RKNN, TensorRT).
Knowledge of Containerization (Docker/Podman) on embedded devices.
Experience reducing OS footprint (Headless optimization, removing GUI/Bloatware).
Familiarity with secure boot and disk encryption (LUKS).
What You Will Be Doing Day-to-Day
Optimizing the Boot: Stripping down a standard Ubuntu image to remove the Desktop Environment (GNOME) and freeing up RAM for the AI models.
Enabling the Connection: Configuring the USB-C port so that when a user plugs the device into a Windows PC, it immediately recognizes it as a "Local Network" without installing drivers.
Connecting the Brain: Writing the glue code that ensures the rkllama AI server has access to the NPU hardware immediately upon boot.
Why Join Us?
Work on Sovereign AI Hardware critical for National Defense.
Move beyond standard "Web Apps" and work on the intersection of Hardware, Linux Kernel, and Generative AI.
Direct impact: The code you write will be deployed in air-gapped, mission-critical environments.
Please record a short <1 min video below. We care less about the camera, lighting, etc and more about what you say.