Summary
Roman Lebedev is a computer vision team lead with a decade of engineering experience and a strong academic foundation in applied mathematics and computer science from ITMO University. He has led cross-functional teams at SberDevices to build real-time ML engines for client-side video/audio processing, server-side neural audio denoisers, and tooling for on-device CI and dataset collection, bringing research-grade models into production. Prior roles include core deep learning work on detection, segmentation and OCR for vehicle surveillance and research on Siamese tracking at JetBrains, reflecting a balance of applied R&D and product delivery. Known for managing ambiguity under tight timelines, he combines backend, desktop and mobile expertise with hands-on model lifecycle ownership and a detail-oriented, continuous-learning mindset. An unspoken strength is his ability to translate research insights into pragmatic engineering (e.g., improving camera-tracking heuristics) that measurably improves user-facing experiences.
10 years of coding experience
5 years of employment as a software developer
Master's degree, Applied Mathematics and Computer Science, Master's degree, Applied Mathematics and Computer Science at ITMO University
Русский, Английский