Aleksei Karpov Email & Phone Number

Vancouver, British Columbia, Canada

Researched self-supervised learning methods for monocular depth estimation.Developed novel neural network models combining vision transformers (ViT) and convolution (CNN) architectures for monocular depth estimation.Published results in the IEEE International Symposium on Mixed and Augmented Reality (ISMAR).Designed, trained and tested machine learning systems for self-supervised depth estimation. Built data pipelines.

Moscow, Moscow City, Russia

Developed deep learning models using PyTorch for personal speech enhancement and speech separation tasks from initial requirements to product implementation. Prepared speech dataset using custom Digital Signal Processing and neural network algorithms for data cleaning and augmentation of speech data from LibriLight and VoxCeleb datasets.Optimized models for real-time inference in smartphones and laptops.Created a multi-threaded Android app to test model real-world environment, collect test data and perform error analysis.Collaborated with the deployment team to integrate models into MateBook X Pro 2022.

Master'S Degree, Computer Science

Bachelor'S Degree, Applied Mathematics And Physics