AI Engineer

Senior Computer Vision / GPU Engineer (C++ / C#): real-time video processing, GPU acceleration, high-load orchestration

We are looking for an engineer for a project where one system runs thousands of instances at the same time. The system reads the screen in real time, processes video streams on the GPU, and makes decisions based on algorithmic strategies.

Please note: this is not web development. The work is on the level of C++/C#, GPU memory, and low-level optimization. We do not consider candidates without real hands-on experience at this level.

Responsibilities

• Build and optimize real-time video and image processing pipelines (HD, 30+ fps, dozens of instances per node)
• Train and run neural network detectors: object detection, OCR, tracking, pattern matching
• Speed up work on the GPU: CUDA, NPP, TensorRT, Video Codec SDK
• Multithreading and parallel computing: TBB, process pools, synchronization between instances
• Orchestrate and monitor a large number of instances, work with queues and inter-process communication
• Profile the system and find bottlenecks: CPU, GPU, I/O
• Algorithmic tasks: combinatorial optimization, integer linear programming (OR-Tools / COIN-OR), time series analysis

Requirements

• C++ and/or C# on production level, Python for the ML part
• OpenCV and neural network detectors (YOLO family, EasyOCR / Tesseract / Nvidia OCR)
• Real experience with GPU acceleration: not only calling a ready framework, but also profiling, memory management, and batch strategies
• PostgreSQL, Docker, message brokers, gRPC or REST
• English from Intermediate (B1)

Nice to have

• Projective geometry: homography, perspective correction, camera calibration
• Full MLOps cycle: data labeling → training → deployment
• Algo trading: ARIMA, regression analysis, RANSAC
• Experience with systems that handle thousands of parallel sessions