Algorithm development for embedded devices impose the challenge of developing a robust algorithm on a limited resources device (in terms of memory capacity, CPU frequency, power consumption & Real-time requirements).Therefore, I use a 4 phase approach:▪Phase 1: Log sensors' data to be analyzed offline. ▪Phase 2: Develop optimized algorithm on a PC framework (e.g. Matlab, TensorFlow).▪Phase 3: Use variety of arithmetic techniques to simplify algorithm complexity, yet maintain desired performances. ▪Phase 4: Convert & implement algorithm on edge device & verify results against the PC framework.I specialized at extracting meaningful insights from sensors' data using signal processing, statistics, machine learning & deep learning. Another aspect which I am focus on is designing an efficient & robust software architecture,I find it as a crucial key in order to meet project timeline while reducing the chances for hidden bugs. Therefore, I first learn the hardware in-depth in order to maximize its' usage and offload software tasks,then I design a modular architecture with defined requirements and clear interfaces.A clear design assist SW engineers to implement faster & precisely.I am part of a dedicated engineering team.