Firmware Architect – RISC-V MCU & Embedded ML

Build the firmware platform for a 32-bit RISC-V MCU family, and help shape the silicon it runs on.

This role spans boot code, drivers, SDKs, real-time control and embedded ML. You will own how those pieces fit together, then use measurements from real workloads to influence future instructions and accelerators.

The devices target precision agriculture and industrial control. Applications include motor and pump anomaly detection, vibration-based predictive maintenance, precision sensing and adaptive control, within tight memory and timing budgets.

Adelaide, South Australia | Full-time, permanent | Hybrid, with lab attendance as required

The role

You will own the firmware architecture from pre-silicon development to production SDKs and customer adoption. You will provide technical leadership to 4–8 firmware, driver and tools engineers, with no direct reports.

• Define boot ROM, bootloaders, startup, runtime, HAL, low-level drivers and SDK structure.
• Set the bare-metal and RTOS approach, including FreeRTOS or Zephyr, deterministic interrupt handling, CLIC configuration and context switching.
• Architect secure boot, firmware updates, key handling and debug lock alongside the silicon security features.
• Own the developer toolchain, including GCC/LLVM, debug and trace, IDE integration, examples and documentation.
• Architect motor-control and sensing libraries: FOC, sensorless observers, 20–100 kHz control loops, synchronised PWM/ADC operation and feedback interfaces.
• Define an on-device ML stack covering model import, int8/int4 quantisation, runtime execution and SRAM/flash planning.
• Optimise neural-network and signal-processing kernels using RVV 1.0, embedded Zve32x/Zve32f profiles and packed-SIMD/DSP instructions.
• Integrate suitable ML tools and libraries, such as TFLite Micro/LiteRT, TVM and IREE, with reference applications customers can use.
• Profile workloads and work with the ASIC Architect on custom instructions, dot-product or matrix operations and potential ML accelerators.
• Develop against FPGA prototypes, simulators and virtual platforms, then lead firmware bring-up on first silicon.
• Set coding, review, test automation and CI practices, mentor engineers and support customer performance tuning.

What you will bring

• 10+ years in embedded firmware, including at least four years in an architecture or lead role.
• Ownership of firmware architecture for at least one MCU or SoC family through to production.
• Expert C, strong C++ and assembly capability for performance-critical code.
• Strong knowledge of RISC-V privileged architecture, CLIC/PLIC, PMP, calling conventions and GCC/LLVM.
• Hands-on SIMD or vector optimisation, using RVV, Arm Helium/Neon or DSP extensions, including intrinsics and analysis of compiler output.
• Experience deploying quantised neural networks on microcontrollers within constrained memory and latency budgets.
• Real-time motor-control firmware, high-rate ADC/PWM handling and sensor signal processing experience.
• Understanding of RTOS internals, bootloaders and secure firmware updates.
• A relevant engineering or computer science degree.

Also valuable

RVV/Zve32 kernel development; ML compilers or custom operators; custom RISC-V instructions; FOC and sensorless control tuning; time-series ML; CAN FD, ISOBUS, EtherCAT or Modbus; Python tooling; and open-source embedded or ML contributions.

Interested? Apply with your CV and a brief example of firmware architecture you owned, including a performance, memory or real-time constraint you resolved.