Khurshid Juraev

Machine Learning Engineer

cartouch.ai•  June 2022 - Present

Used ML models to automate photoshopping services using deep learning. Several car selling platforms in the UK were in need of car background removal, window tilting, shadow extension and background replacement services. I developed, tested, deployed deep learning based image processing tools to serve these customers. ● Automated more than 95% of the tasks using instance segmentation as well as inpainting models. ● Teaching deep learning to talented and selected students, usually olympiad winners, as part of the internship program. Gathered and led a unique team of self-taught, self-organized, young deep learning engineers in Tashkent. This team still exists and is growing. ● https://cartouch.ai

Machine Learning Engineer Intern

Alemira•  June 2022 - August 2022

● Conducted survey on pose estimation models. ● Data preparation for video annotation models. ● Conducting survey on spatio-temporal learning models.

Software Engineer

Finnet Technologies•  October 2019 - November 2020

Developed a sports exercise assessment library(Praaktis SDK) for android/iOS using Google MLKit pose estimation software. Directly influenced the architectural improvement of the library by actively doing R&D practices, experimenting and design thinking. Wrote analyzers to analyze particular motions of users such as squats and lunges. Conducted a weekly meetup with my coworkers on topics of Algorithms and Data Structures. Key Accomplishments: ● Reduced monthly costs by thousands of USD per a cluster of ten users. At some point, Praaktis SDK was based on a Pose Detector tool hosted on the cloud. I did move the pose detection logic to mobile phones and reduced networking and hosting expenses. ● Achieved high-accuracy analyzers with pseudo-real-time assessment report and feedback. Developed strategies for building analyzers for dozens of exercises types of which range from gym, physiotherapy and martial arts. ● http://gympraaktis.com

Labaratory Assistant

Tashkent University of Information Technologies•  October 2018 - June 2019

● Taught freshmen to solve competitive programming problems, to apply common techniques and algorithms. ● Implemented several computer vision modules such as text-image binarization methods, segmentation of texts to lines.

Software Engineer

Phoenix LLC•  March 2018 - July 2018

● Integrated CRM system to the company’s e-commerce web platform.

National University of Singapore (NUS), Singapore

Computer Science, MS•  August 2021 - January 2023

It is a degree with AI specialization. Courses taken: * CS5234 - Algorithms at Scale * CS5242 - Neural Networks and Deep Learning * CS5340 - Uncertainty Modeling in AI * IS5009 - Topics in Financial Technology Solutions * IS5116 - Digital Entrepreneurship * CS5446 - AI Planning and Decision Making Courses Audited: *CS5238 - Advanced Combinatorial Methods in Bioinformatics *MA5249 - Stochastic Processes and Algorithms *CS6213 - Network, Crowds and Markets Dissertation Project: ADSOLVER: AN ADAPTIVE SOLVER FOR COMPREHENSIVE CAPTCHAS IN THE WILD