Valiantsin Yauseyenka Email & Phone Number

Warsaw, Mazowieckie, Poland

- Designed and implemented end-to-end solution for automated drone orthophotos analysis which speeded up 10 times labeling routine- Developed object detection and segmentation algorithms for large scale images with SOTA segmentation techniques to hit and exceed human level quality- Led text-recognition solution for knowledge extraction from real world data heavily using MMDetection project- Delivered text similarity search service based on LLM and FastAPI backend- Supervised data mining and labeling processes for object detection, text-detection and text-recognition tasks

Ташкент, Ташкентская Область, Узбекистан

- Developed suite of segmentation, detection and classification models for intra-operative surgical applications using Pytorch and Keras/Tensorflow and deployed them with ONNX and TensorRT- Workerd on AR-integrated 3D segmentation for complex cases with MONAI and Torchvision neural networks architectures- Led image registration workflow usign classic SImpleElastix approaches and complex 3d neural networks for medical scans- Speed up CT/MRI scans annotation using segmentation neural networks- Developed and deployed video analyzing system on Azure ML platform via custom pipelines- Worked as a lecturer ML on courses

Нур-Султан, Казахстан

- Optimized NN models using OpenVINO and TensorRT for efficient inference- Developed efficient version of Non Maximum Suppression algorithm on C++- Researched and developed object tracking system via C++ 17, Eigen, OpenCV- Applied custom object-detection algorithms in PyTorch and ONNX for realtime video analytics

Минск, Республика Беларусь

- Trained pose-estimation and keypoint detection models using Keras/Tensorflow frameworks- Designed autolabeling tool for time-series with Scikit-learn package- Developed digtal signal processing pipeline on python and optimized it with C++ Eigen framework- Optimized and deployed time-series RNN models on iPhone using CoreML