Computer Vision Researcher with a proven track record in delivering high-performance vision systems in diverse domains. My expertise covers the whole machine learning project lifecycle: from managing data gathering and annotation, to developing and training state-of-the-art approaches and optimizing them for real-world usage.Proficiency with classical and deep learning based computer vision approaches:- feature extraction, object detection, and recognition tasks;- worked with CNN and Vision Transformer based models:ResNet, ResNeXt, EfficientNet, YOLO, ViT, ConViT, CLIP, etc.;- highly experienced with handling imbalanced data: focal loss, SAM, Yogi optimizer, augmentation techniques, synthetic data generation and other rebalancing methods;- dataset preparation: cleaning, outlier/anomaly detection, labeling optimization;- extensive experience with classical and geometry-based algorithm development;- applying state-of-the-art methods to get significant performance gains.Additional ML/DL skills:- generative and sequential approaches: RNN, LSTM, GANs;- similarity metrics, embedding optimization, and semantic space search with FAISS;- report generation, data exploration, and building visualizations;- basic experience with LLMs, Prompt Engineering, and RAG: GPT-4, LangChain.