Embedded Linux Engineer (Yocto Specialist)
Confirmed on the employer's own hiring board on Nov 10, 2026. First seen by Alion on Nov 10, 2026. Genesis AI scores A on the Alion truth index.
About Genesis
Genesis is a physical AI lab and full-stack robotics company building general-purpose robots, with teams in Paris and the Bay Area. Our people have built some of the most widely used platforms in robotics research including the open-source Genesis simulator, pioneered generative simulation and led foundation-model work at some of the world's leading AI labs. We own the whole stack: hardware, embedded systems, simulation, data and robot foundation models.
The Role
Every Genesis robot runs on a custom jetson thor compute board and every piece of software we write: control, perception, learned policies, depends on the Linux image underneath it. We are looking for a senior engineer to own that image end to end: build it, ship it to the robot, keep it updatable, and make it easy for the rest of the team to compile and deploy on top of it.
This is an in-house ownership role. You will be the reference for embedded Linux at Genesis, working closely with our embedded, robotics and application-software engineers in Paris and with our hardware partners. The job is to provide a reliable, well maintained and up to date os image, customized to our needs.
What You'll Do
Build, own and maintain the production Yocto/OpenEmbedded distribution for our robot's custom board: BSP, kernel, device tree, rootfs, toolchain and custom layers.
Bring up new hardware on NVIDIA Jetson-class platforms, integrating drivers and peripherals. Including hands-on debugging of components such as Wi-Fi, cameras, audio devices and buses on the physical robot.
Own the full image lifecycle: reproducible builds, versioned releases, secure and robust over-the-air (OTA) updates, and long-term maintenance as hardware and software evolve.
Deliver a development kit and workflow (SDK, cross-compilation, deployment tooling) so engineers can build and ship their software onto the robot quickly.
Integrate GPU and AI-accelerator inference stacks so that our models run efficiently on the robot.
Speed up the team's build and CI pipeline for image generation and validation. Builds are long, so caching and automation matter, as well as removing unnecessary dependencies.
Document the system and make it approachable, so the image is never again a black box that only one person (or contractor) understands.
Your First Three Months
Build a custom Yocto image for the robot's custom boards.
Flash, boot and validate it on real hardware, working through driver and integration issues on-site.
Set up the cross compilation workflow, caching, CI automations to streamline the robotics team work
What You'll Bring
Passion for your craft and demonstrated excellence in embedded Linux engineering.
5+ years of hands-on embedded Linux experience, including shipping at least one Yocto-based image to production not only a prototype.
Proven ownership of an image through its lifecycle: maintenance, release management and remote/OTA updates.
Deep experience with Yocto internals: layers, recipes, BitBake, machine and distro configuration, SDK generation.
Strong board bring-up and low-level debugging skills: kernel configuration, device trees, drivers, boot chain.
Hands-on experience with the NVIDIA Jetson family (JetPack / L4T, meta-tegra) or comparable embedded GPU platforms.
Knowledge C++ and Python packaging systems and how to deal with cross-compilation.
The ability to explain how your work fits into the wider build, development and deployment system.
Exceptional ownership and initiative, with the judgment to choose workable solutions for our current stage over the theoretically ideal one.
Bonus
Experience integrating GPU / AI-accelerator inference (CUDA, TensorRT or similar) on embedded targets.
Background in robotics, drones, connected devices or other products with custom hardware.
Experience working with chip vendors or ODMs.
Familiarity with secure boot, real-time kernel tuning (PREEMPT_RT) or OTA frameworks such as Mender, RAUC or SWUpdate.
French and English. Mandarin is a plus.
You'll Thrive Here If You
Enjoy being the go-to person for a critical system and making it easier for others to use.
Unblock teammates' requests quickly and pragmatically, and know when "good enough to ship" is the right answer.
Stay patient and methodical when a single build takes two hours and a bug only shows up on hardware.
Like startup flexibility: changing priorities, short feedback loops and work that ranges from kernel to tooling.