> Markdown version of [/jobs/ext/4699738-senior-embedded-software-engineer](https://www.wearedevelopers.com/jobs/ext/4699738-senior-embedded-software-engineer). Every page supports `.md` or `Accept: text/markdown`. Links point to the HTML versions so they work for humans too. Agent guide: [/agents.md](https://www.wearedevelopers.com/agents.md). --- # Senior Embedded Software Engineer - **Company:** Toumetis - **Location:** Bristol, United Kingdom (Remote available) - **Experience:** Expert - **Salary:** £60,000 - £70,000 - **Contract:** Permanent contract - **Skills:** Adobe Flash, Artificial Intelligence, ARM Architecture, C++ (Programming Language), Software Code Optimization, Embedded Software, Firmware, Python (Programming Language), Real-Time Operating Systems, Tensorflow, Reduced Instruction Set Computing, Signal Processing, PIC Microcontroller, Pytorch, Hardware Infrastructure - **Published:** November 13, 2026 - **Apply:** https://www.adzuna.co.uk/jobs/details/5962892216 ## About the Role * Languages: Expert proficiency in C/C++/Rust (for the device) and Python (for the model pipeline). * Frameworks: Deep experience with TinyML tools (TFLite Micro, CMSIS-NN, or STM32Cube.AI). * Hardware Knowledge: Familiarity with ARM Cortex-M series or RISC-V architectures. * Mathematical Foundation: Strong understanding of linear algebra and signal processing (FFTs, digital filters). ## Description * Model Optimization: Convert high-level Python/TensorFlow/PyTorch models into optimized C++, Rust or Flatbuffer formats using TensorFlow Lite for Microcontrollers or Edge Impulse. * Resource Management: Implement techniques like 8-bit quantization, weight pruning, and knowledge distillation to ensure models run within <256KB of RAM and minimal Flash storage. * Firmware Integration: Collaborate with Embedded Engineers to integrate inference engines into the device's firmware (RTOS-based) without disrupting primary metrology functions. * Signal Pre-processing: Develop efficient Digital Signal Processing (DSP) pipelines to clean raw voltage/current data before it reaches the AI model.