Raman Rusak Email & Phone Number

Cyprus

Leads the design and development of C/C++ software architecture, signal processing and machine learning algorithms for resource-constrained Edge devices(MCU, sensors) and the AutoML/TinyML web platform.

Minsk, Belarus

• Implemented a production-grade Anomaly Detection model for a Cortex-M4F based HVAC system.• Developed C++ tinyML library with edge-learning functionality to prevent Dataset shift problems.• Preparation of materials for marketing and scientific events like tinyML Summit.• Developed LTE-CATM1 network stack and drivers for working with Sierra Wireless modem.• Implemented LoRaWan Class-C network features for rental property management device.Engaged in the development of ML models and wireless stacks for clients in the field of Smart Access Control and HVAC systems. And also involved in the enhancement of the company's TinyML competencies, R&D and business proposals.

Minsk, Belarus

• Working closely with Machine Learning team, provide Predictive Maintenance and Keyword Spotting features on Cortex-M4F.• Implemeted different DSP algorithms: FFT, MFCC, DCT and more.• Improve performance of ML algorithm using ARM SIMD and vector instructions.• Developed new STM32Cube Expansion Package with Vibration Analysis features.• Wrote various design documents which describe the firmware architecture, implementation and interactions within the various subsystems.• Porting Azure and AWS IoT Device SDK and wrote services to work with server backend.• Developed LoRa Class-A Node Device as part of Actility ThingPark™ ecosystem.• Performed power consumption analysis and developed firmware for lowest battery consumption.• Wrote software for testing the device in production, helped set up production at the factory.• Design and implement Board Support Package for IoT-Node device.Developed SMARTEDGE AGILE IoT solution with end-to-end security stack, machine learning algorithms for Predictive Maintenance, Anomaly detection and Motion Recognition.

Minsk, Belarus

• Optimize Machine Learning algorithms for performance improving on Cortex-M3/M4F• Implement low level device drivers for multiple MEMS sensors.• Developed drivers and Adaptive Kalman Filter Positioning Algorithm for GNSS receiver.• Developed test tool for automated in-target software embedded testing.• Developed firmware for Li-ion battery state of charge/discharge estimation and power management features.• Implemented LoRa and (G)FSK M2M communication stack.• Assisted in hardware design, resulting in significantly reducing debug and integration time.Working on PIQ and Babolat POP smart wearable sensors with motion recognition algorithms that analyze sports movements and provide athletes with statistics and tips for improvements.

Minsk, Belarus

• Maintenance, installation, dismantling various medical equipment: EKG / ECG machines, diagnostic ultrasound systems, electrosurgical devices, medical air disinfection & purification systems.• Adjustment, repair, technical support of medical equipment.• Performed analyses and evaluation on data samples and documented information gathered during inspections of medical equipment.