Summary
Stanislav Dereka is a Senior ML Engineer with eight years of experience specializing in computer vision, probabilistic embeddings, and robust model uncertainty methods. He has driven production-ready CV and multimodal projects at Tinkoff and VK, published multiple peer-reviewed papers on image retrieval and OOD detection, and mentors junior engineers and interns. His work spans the research–production gap: from state-of-the-art pipelines (deepspeed, Hugging Face, distributed training) to pragmatic gains like a 4.6% ROC-AUC lift while cutting dataset size by 60% and quantisation strategies that doubled inference speed with minimal accuracy loss. Currently pursuing a PhD at MIPT, he pairs academic rigor with hands-on deployment experience (Docker, ONNX, W&B), and has a track record of turning CV research into measurable product impact.
8 years of coding experience
6 years of employment as a software developer
PhD Artificial Intelligence, PhD Artificial Intelligence at Moscow Institute of Physics and Technology (State University) (MIPT)
English, Russian, French