At Lobachevsky Research Institute (LRI), Huawei, I modernize the Mesa graphics driver with dithering for BC1 compression and ISPC-based BC7 support. I also build Python tooling to evaluate rendering quality and analyze performance.
I developed a high-performance CPU inference library for ROI detection using ONNX Runtime, tiled inference, and OpenMP. I also designed and maintain an end-to-end detection-model pipeline, from dataset preparation and Ultralytics YOLO training through ONNX export and benchmarking.
As a private tutor in mathematics and computer science, I prepare high school students for the Unified State Exam. I teach topics including algorithms, data structures, and computational geometry.
My projects include a YOLO training and deployment pipeline, a smart voice assistant for drivers, and a convex-polygon intersection area calculator. I’m studying Applied Mathematics and Information Science at HSE University.
