Taras Hevlich

Architected and developed a comprehensive Python-based ecosystem and a BDD framework for IP cameras and Network Video Recorders (NVRs), heavily utilizing asyncio, queues, and multiprocessing for parallel task execution. Developed a core interaction library implementing diverse connection protocols including gRPC, WebRTC, RTSP, SSH, and UART for robust device configuration and log analysis. Optimized high-load WebRTC stream processing by reverse-engineering and patching the aiortc library to store encoded media packets directly in RAM, deferring CPU-intensive decoding tasks to post-stream execution. Engineered an image and video analysis module integrating PyAV, OpenCV, NumPy, and SciPy for on-the-fly RTSP video stream analysis and image validation. Designed and deployed standalone FastAPI monolithic applications packaged via Docker into local executables. These tools orchestrated complex multi-device environments involving proprietary hubs, smart sockets, and IoT cameras to validate of

Expanded an existing E2E automation framework (based on pytest and asyncio) dedicated to the validation of SmartHome devices over proprietary sub-GHz radio protocols. Developed highly resilient communication algorithms with custom retry strategies, timeouts, and byte-sequence validation to ensure stable interaction over noisy RF environments and physical interfaces (UART, TCP). Automated hardware testing by integrating programmable power blocks into physical test stands, controlling both via UART and TCP APIs to simulate overcurrent and overvoltage scenarios. Orchestrated concurrent execution flows to dispatch hardware commands and simultaneously validate the real-time reactions and protection mechanisms of embedded devices.