Andrey Kuznetsov Email & Phone Number

Leading multimodal and generative AI research teams in the following fields: multimodal models for vision, audio, text and other modalities; multiagent LLM algorithms and frameworks; modalities fusion approaches; intrinsic embedding attributes investigation; generative AI for text-to-image and text-to-video synthesis; attacks on LLMs and multimodal models.FusionBrain Laboratory consists of 5 core research teams:1. Multimodal and multiagent models design2. Large models architectural insights3. AGI for robotics4. Video Generative AI5. Controllable Generative AI

Lead and plan research in computer vision and multimodal fields of research, develop AI-based software, publish papers in Q1 journals and A/A* conference proceedings, achieve SoTA quality results. Leading a team which is responsible for: 1) GHOST - one shot face swap model to transfer a face from a single image to image or video (SoTA) 2) Kandinsky - an autoregressive text-to-image generative model 3) Kandinsky 2.0 - a diffusion-based text-to-image generative model 4) Kandinsky 2.1 - a diffusion-based text-to-image generative model with strong quality enhancement (open-source SoTA) 5) GIGACHAT - a multimodal chat GPT-like model 6) RUDOLPH - a multimodal multitask single autoregressive architecture 7) FaceSearch - a fast video analysis tool for face reidentification

Samara, Samara, Russia

Develop AI solutions for computer vision using machine learning and multimedia data processing methods. Actively use deep learning methods to achieve results. Conduct research in the field of remote sensing data analysis, digital forgery detection, computer vision, machine learning and image processing.Supervise several RFBR grants on the topic of digital forgery detection for images and videos, including the use of deep learning methods.Develop methods, algorithms and software using C # and Python programming languages. Developed several commercial products using ASP.NET Core.Publish scientific papers (approx. 40 in Scopus and Web of Science databases). Teach the author's course "Technology of Designing Secure Distributed Applications" at the Department of Geoinformatics and Information Security. Give lectures on science popularization for applicants and students.

Developed more than 20 projects on software development and computer vision. Top Rated developer.

Самара

Разработка программного обеспечения в области геоинформационных систем и анализа данных (цифровые изображения, видео и т.д.)

Разработка программного обеспечения с использованием С++, .NET и AJAX. Приложение для кинотеатров, биллинговые системы

Specialist, Mathematician, System Developer, 5.0

Кандидат Технических Наук, 05.13.17 "Теоретические Основы Информатики", 5.0

Activities and Societies: Методы обнаружения искусственных искажений данных дистанционного зондирования ЗемлиPhD in Computer Vision.