Experience

Rubius

• IT Services and IT Consulting
• 1 - 100 Employee

• Senior Machine Learning Developer

  • Apr 2022 - Present

  In my current role at Rubius Gulf, I've been working on a variety of projects:1. Traffic Analytics System:- Implemented a real-time person re-identification module using Torchreid for feature extraction, which improved the Traffic Analytics System's metrics.- Adapted previous work from the traffic analytics system to develop a multi-camera tracker for sport game analysis, which was well-received by the client.- Successfully pitched a video analytics project at Astana Hub, aimed at optimizing shop location selection, leading to project approval and recognition.2. Other Projects:- Rectified a broken integration with downstream processes by resolving bugs and tailoring output format in a C++ project for point cloud data ground segmentation and oct-tree construction.- Contributed to a C# project implementation of algorithms, including dynamic programming and geometry techniques, to address resource limitations and accelerate project development.- Frequently conducted analysis of custom development requests, including the formulation of technical proposals, task assessments, and estimation of expected labor costs. Show less

• Senior Machine Learning Developer

  • Jan 2022 - Apr 2022

  As a Senior Machine Learning Developer at Rubius, I led a team of 3 developers to enhance the precision of prediction for Supply Chain Management by 2 times.My responsibilities included designing and maintaining a Python-based data pipeline with pandas for Supply Planning, which led to improved forecasting accuracy.In addition to my technical work, I actively participated in maintained ongoing communication with client to deliver progress updates and presentations.

• Machine Learning Developer

  • Oct 2018 - Jan 2022

  In my role at Rubius, I contributed to several key projects:1. Traffic Analytics System:- Designed a multiprocessing architecture, boosting the performance of a real-time video analytics solution and resulting in a 2.5x increase in average FPS.- Developed a real-time online multiple object tracker integrating various Neural Network models, achieving a satisfactory MOTA of 0.38 on a custom dataset while maintaining high FPS.- Reduced counting error rate by 2.45 times compared to human observers by integrating the object tracker with a line-crossing detection algorithm.- Ensured the stability of client's video streams and prevented data losses by developing a Python solution for restreaming and health-checks of unstable RTSP video streams.- Enhanced data preparation efficiency for training various models by developing auto-annotation and data processing tools.2. Medical Decision Support Module:- Developed a decision support module featuring a cascade classifier that emulated a doctor's decision-making process, achieving a 98.4% cross-validation accuracy on balanced data.- Improved data accuracy for the decision support module by analyzing device data, identifying, and rectifying discrepancies using Python tools.3. Other Projects:- Optimized quality control processes by applying real-time computer vision techniques to analyze conveyor camera footage and detect defects.- Enhanced decision tree algorithms' robustness by building a Neural Network-based solution using TensorFlow.- Accelerated audio data processing 10 times by developing C++ wrappers and a Python module.- Supervised a team of annotators and devised effective annotation strategies, achieving strong model performance in a Tacotron-2 speech synthesis project. Show less



Tomsk State University's Math Center

• Tomsk, Russia

• Teacher

  • Mar 2021 - Aug 2021

  Taught "Introduction to Competitive Programming with C++ and Basic Algorithms" at Tomsk State University's Regional Scientific and Educational Mathematical Center, receiving positive feedback. Taught "Introduction to Competitive Programming with C++ and Basic Algorithms" at Tomsk State University's Regional Scientific and Educational Mathematical Center, receiving positive feedback.



Tomsk Polytechnic University

• Russian Federation
• Higher Education
• 1 - 100 Employee

• Teacher

  • Sep 2019 - Jun 2021

  I started as a mentor for a Competitive Programming beginner's group, presenting lectures and practices, sometimes managing up to 30 participants. Later, took a lead of the competitive programming initiatives at the university, monitoring all students, achieving regional results, and successfully training a successor. Also contributed to the organization of city-wide programming competitions. I started as a mentor for a Competitive Programming beginner's group, presenting lectures and practices, sometimes managing up to 30 participants. Later, took a lead of the competitive programming initiatives at the university, monitoring all students, achieving regional results, and successfully training a successor. Also contributed to the organization of city-wide programming competitions.



Rubius

• IT Services and IT Consulting
• 1 - 100 Employee

• Machine Learning Intern

  • Jun 2018 - Aug 2018

  During my internship at Rubius, I achieved a 95% accuracy rate in mapping calls within anonymized ambulance data utilizing text processing techniques and tools like Python, Pandas, and Jupyter Notebook. This data was later used by the team for visualization, analysis, and call prediction. One of the significant challenges was working with error-prone and incomplete data. During a Demo Day presentation, I highlighted these challenges, and as a result, the project received additional support for improving the data gathering process. Show less