About the position
OpenAI is looking for a System Software Engineer to design, implement, and debug the foundational software that enables our products. You'll work across the OS stack-kernel, drivers, and userspace daemons-and own pipelines for building and updating device images. Your work will span performance tuning, security hardening, and deep system debugging. We're looking for engineers who thrive in complex, cross-disciplinary environments, can take ownership from concept to release, and are passionate about shipping robust, production-grade systems.
Responsibilities
- Design, implement, and debug kernel modules, device drivers, and userspace daemons
- Configure and maintain the OS: kernel, initramfs, init services, udev rules, networking, and security hardening
- Build and operate image/OTA pipelines (Yocto/Buildroot, package management, A/B updates, dm-verity)
- Instrument and analyze performance with perf, eBPF, ftrace; optimize CPU, memory, I/O, and power usage
- Own system observability and reliability: logging, crash capture, watchdogs, and field diagnostics
- Collaborate with hardware, camera/vision, and application teams to define interfaces and deliver features
- Establish strong engineering practices: code review, CI, reproducible builds, and release management
- Work with vendors and support system builds in the factory
Requirements
- Deep experience shipping Linux-based products on hardware devices (arm64 or x86)
- Proficient in C/C++ and comfortable with embedded Linux internals (drivers, IRQs, power management, filesystems, networking)
- Strong systems debugging skills with gdb, strace, perf/eBPF, jtag, and kernel logs
- Understand kernel device tree overlays, I2S, I2C, SPI, CSI/DSI
- Fluent in userspace foundations: systemd, D-Bus, sockets/IPC, udev, networking tools, shell/Python
- Experience building OS images and packages (Yocto/Buildroot, Debian/RPM, Docker) and OTA updates
Nice-to-haves
- Upstream contributions to Linux or major subsystems
- Experience with secure boot, platform security, SELinux/AppArmor, TrustZone/OP-TEE
- Graphics/ISP or accelerator driver exposure (GPU, V4L2, media/codec, DMA, shared memory)
- Wireless connectivity stack experience (Wi-Fi/BLE, NetworkManager, wpa_supplicant)
- Production observability at the edge (symbolized crash reports, metrics, remote diagnostics)
- Experience in robotics, mobile, or other resource-constrained devices
- Familiarity with on-device ML accelerators, audio pipelines (ALSA, DSP, WebRTC, gstreamer)
- Knowledge of protobuf and other serialization formats
