Ruslan Moldagazyyev

Graduate Research Assistant

ISSAI

Institute of Smart Systems and Artificial Intelligence at Nazarbayev University.

Collects multilingual synthetic speech data and researches an open-source speech-to-text model for Kazakh, Russian and English.

• Collecting synthetic datasets with emotional prompts and text for Kazakh, Russian and English. • Conducting research on building a multilingual open-source speech-to-text model for Kazakh, Russian, and English under the supervision of ISSAI Lead Data Scientist, Askat Kuzdeuov.

AI Engineer

Victory Technology

Built speech transcription, testimony contradiction detection and question-generation modules for a law-enforcement system now in pilot testing.

• Contributed to the development of an AI system to assist law enforcement by automating investigative tasks and reducing manual workload. • Focused on implementing modules for speech-to-text transcription, contradiction detection in witness testimonies, and generation of clarifying questions for interrogations. • The system is currently in pilot testing within police departments in Astana and the Akmola region.

Machine Learning Engineer

Nazarbayev University

A research university in Astana, Kazakhstan.

Developed human pose-estimation algorithms for an AI-based football motion analysis project.

• Developed algorithms for human pose estimation under the supervision of Professor Sultangali Arzykulov in the project "AI-Based Human Motion Analysis in Football."

Graduate Research Assistant

Nazarbayev University

A research university in Astana, Kazakhstan.

Compared CNN architectures for solar-panel contamination detection; the best model achieved 97% accuracy across four classes.

• Conducting research on "Solar Panel Contamination Detection and Classification Using Machine Learning" supervised by Professor Mehdi Bagheri and Professor Amin Zollanvari. • This research conducted a comparative analysis of VGG, ResNet, EfficientNetB2, and ConvNeXt Tiny architectures for solar panel contamination detection, achieving 97% four-class accuracy with the best-performing model.

Computer Vision Intern

Cleverest Technologies

Developed YOLOv8-based weapon and trash detection models and contributed to commercial pilot projects.

• Interned as a Computer Vision and ML Engineer at Cleverest Technology, specializing in developing YOLOv8-based ML models for weapon and trash detection. • Gained expertise in ML and CV for commercial use and contributed to pilot projects.

MS, Electrical and Computer Engineering

Nazarbayev University

Data Science

Outpeer KZ

BEng, Electrical and Computer Engineering

Nazarbayev University