All Projects

Machine learning approach to automatic PID controller parameter tuning for robotic systems. Developed ML-based method for instant PID tuning from robot parameters (1k simulations). Neural network predicts optimal PID parameters from robot physical characteristics in milliseconds, achieving 78-90% performance improvement over traditional methods. Generalizes across different robot configurations; simulation-based, hardware validation planned.

Linear actuator control system using Arduino and PID controller for tensegrity robotics. Development applied principles of tensegrity (tensile integrity) - creating stable structures through force balance.

AI-powered logging platform that fixes production issues in seconds with one click. Tiny SDK sends structured logs enabling instant semantic search, debugging, and AI-generated fixes.

Arduino-based stepper motor system with custom PID for adaptive resistance, designed for post-stroke & ASD patients. Mechatronic rehab system for motor-function recovery after neuromotor impairments.

Platform for creating AI agents for businesses: automating customer communication in messengers, CRM integration, and freeing up time for important tasks.

Product Hunt-style launch platform for Kazakh founders and startups. Platform for Kazakhstani founders to publish products, get feedback, and find support from the local community.

Open-source core motion library developed over years and left open for students and the community. Essential functions: line following, gyro-stabilized movement, cosine-smoothed ramps, encoder-based distance control. Created to help WRO teams focus on strategy instead of basic motion code.

Pet project evolved into practical solution - comprehensive document workflow management system for local company. Successfully implemented and actively used for streamlining document approval processes.

Professional computer vision system for automatic detection and classification of car damage using YOLO and machine learning. Two-stage system: damage detection + severity classification.

Pet project: Web application for creating and studying flashcards with artificial intelligence integration. Helps users learn through interactive flashcards with AI-generated content.

First experimental project - smart firefighter costume with integrated safety features. Extensive work and analysis went into developing this early prototype.