I am an engineer and a researcher passionate about computer vision, machine learning and deep learning with a strong background in computer science and math.Currently working on medical image processing helping radiologists making correct diagnosis. I am leading a team bulding accurate image processing solutions for over 100'000 patients. My 7 year working exeprience included over 20 success AI projects from scratch to production. A publication at top computer vision conference CVPR 2023. Master's degree in Mathematics and Computer Science from Lomonosov Moscow State University.My skills: * 2D and 3D computer vision* Segmentation, detection, recognition* Neural Network architecture optimization* Generative AITech Stack: Python, Pytorch, AWS, SQL, Docker, C++I am open for job opportunities and communication.
-
Chief Technology OfficerAi StarCyprus -
Leading Computer Vision EngineerThird Opinion Ai Jan 2021 - PresentБерлин, Германия• Leading a team developing a Head CT hemorrhage segmentation system integrated in 5 commercial clinics. Achieved metrics 0.92 IOU and 0.95 ROC that is x5 better than radiologists markup and x3 better than competitors.• Developed validation modules for Dicom data; implemented custom 3D medical image segmentation network in Pytorch; adapted the model to work on 6 hemorrhage types constantly checking up the results with radiologists.• Reduced costs per client optimizing production pipeline by 40% GPU resources by Ray and Triton frameworks. -
Senior Computer Vision ResearcherSamsung Electronics Jan 2019 - Jan 2021Москва, Россия• Developed a novel approach for compression generative networks that resulted in the article (https://arxiv.org/abs/2304.10306) at the CVPR 2023 conference. The new approach resulted in SOTA in knowledge distillation, neural architecture search and pruning domains.• Made it possible to use GAN models directly on devices for commercial mobile applications. Reduced models memory x8 and inference time x5 by ideas from the last research article. -
Research And Development EngineerVisionlabs Nov 2016 - Jan 2019Москва, Россия• Developed 10 MVPs which grew into new commercial products. Pipeline consisted of web scrapping data by Requests and BeautifulSoup; training segmentation(Unet, MaskRCNN), detection(Yolo) and OCR Pytorch pipelines; UI implemented by Telegram bots and IOS apps by Docker and CoreML libraries.• Created a new version of a Face Anti Spoofing classification system presented in 20 organizations. Improved quality x5 TPR@FPR better than last production by implementing approaches from SOTA articles from scratch. Reduced data costs by 60% by implementing GAN synthetic data generation pipeline.
Leon Kostyushko Education Details
Frequently Asked Questions about Leon Kostyushko
What company does Leon Kostyushko work for?
Leon Kostyushko works for Ai Star
What is Leon Kostyushko's role at the current company?
Leon Kostyushko's current role is Chief Technology Officer.
What schools did Leon Kostyushko attend?
Leon Kostyushko attended Московский Государственный Университет Им. М.в. Ломоносова (Мгу), Московский Государственный Университет Им. М.в. Ломоносова (Мгу).
Free Chrome Extension
Find emails, phones & company data instantly
Aero Online
Your AI prospecting assistant
Select data to include:
0 records × $0.02 per record
Download 750 million emails and 100 million phone numbers
Access emails and phone numbers of over 750 million business users. Instantly download verified profiles using 20+ filters, including location, job title, company, function, and industry.
Start your free trial