Trained Yolov3 neural network model to detect field objects with FPS ~37 (on 6 frames more than Yolov3 Tiny) and accuracy 61% (on 10% more than Yolov3 Tiny). Constructed Robot using EV3-Mindstorms and Arduino. Connected SQL server with Arduino using Local Web (Wi-Fi module), enhanced real-time event handling to ~0.1 seconds (previous time of reaction was 1.6 seconds). Developed program for python server to handle events and send corresponding signals to the robot using Wi-Fi

Short Description: Development of the city-keeper robot for the successful implementation of the smart city system. City-keeper has various functions such as detecting garbage, watering flowers and eliminating fires, which contributes to the sustainable development of smart city systems. This robot is able to recognize various objects thanks to computer vision and a neural network. Our presentation (Click)

Designed model of wireless charging table. Investigated method of magnetic induction. Constructed system of wireless charging. Achieved charging 3000 mAh battery in 24 hours on distance 2 cm

Short Description: To date, most fires occur due to spontaneous combustion of power supplies of equipment (which is confirmed by statistical data), as a result of prolonged load or power surges. In addition to fires, this can lead to a banal breakdown of equipment, which leads to unnecessary costs. The "Innovative Workplace" project aims to solve this problem. It is an automatic system for charging and powering gadgets using wireless energy transfer technologies. This allows you to get rid of potential sources of fire. In addition, the project improves the ergonomics of modern workplaces, freeing up space from wires and power supplies on tables, thereby increasing the area of the free working surface. It is also worth noting that the workplace of the "Innovative Workplace" project will be organized according to all the rules of ergonomics, to take care of the user's comfort and health. Our presentation (Click)

Trained Neural Network to recognize cars. Implemented Dijkstra's algorithm to detect the shortest route based on the received data about how many cars are on each road. Connected SQL database with python server in which the data about customers and taxi's availability are stored

Short Description: Taxi ordering system based on AI calculations about busy roads and Dijkstra's algorithm (finding the best route). Taxi on autopilot is a future that already comes on the heels of us, i.e. require operational implementation, as the current Informatization and globalization are possible in a developed, advanced and comfortable transport system – car without a driver; clearing; adapts to cork route; fail-safety and, of course, speed and comfort. Our presentation (Click)

Implemented a voice recognition system through the voice module and achieved 91% accuracy. Introduced the system on I-Seed, received recognition and grant on 100k tenge. Implemented voice recognition system in Nazarbayev Intellectual School in FabLAB (technotoria). Developed house control system through the servo motors, gas detectors, light sensors

Short Description: This project is intended for people with disabilities. In this regard, the goal of my project is to create facilities and improve the lives of people with disabilities by using a voice command home management system. As you know, in many developing countries, including Kazakhstan, there are Smart homes, the main prerogative of which is to save electricity. However, they and all other houses are not designed for people with disabilities. For example, they can't reach the switch because they are constantly in bed, or reach the switch without having hands. Many people can't open a door that opens out, or a window. In short, being unfit creates discomfort and prevents them from interacting freely with the home. So my project is to use a voice-activated home management system that will allow people to make it easier to use the house and save time spent moving around in the house. Our presentation (Click)

Developed a Mobile App with Mixed reality that works with the "Ecology" book showing interactive 3D objects.

Short Description: The project is an Autonomous device that can find and eliminate sources of pollution without human intervention. Based on our research, we can conclude that the amount of pollution reaches critical levels, having a significant impact on health and the environment. There are various projects using controlled robots for cleaning, but they depend on human resources, i.e. they require constant human control, a qualified operator, and round-the-clock surveillance. Our project offers the introduction of Robots based on machine learning in order to solve the problems of human dependence and environmental pollution at the same time.Our presentation (Click)

Trained Yolov3 to detect specified objects (debris) in orbit. Programmed robot to navigate on the debris and catch it using electromagnets. Established connection between python server and robot using local network

Short Description: Nowadays, an active process of space exploration is underway, the world is undergoing rapid development of space technology, which is why pollution problems in the near-Earth atmosphere arise. Space debris represents a significant threat to satellites in orbit, as well as for cargo and manned spacecraft. The fact is that a collision with an object larger than 10 centimeters can lead to the complete destruction of the spacecraft. This paper presents the main causes and solutions to the problem of pollution of near-Earth space, and a new method of cleaning the atmosphere. After conducting this study, we are planning to create a new type of space robots that will be able to work autonomously and clean the near-Earth space from debris. After conducting research, we decided to assemble our first prototype, Orbit, a stand-alone device that uses artificial intelligence to search for space debris that will work on ion engines. When building a prototype, we found that the use of artificial intelligence is a great idea, as it greatly simplifies the process of searching for garbage and eliminates a very important minus: the human factor. We built a special platform where we conducted experiments and trained our neural network. The results of this study can be applied in the construction of a full-scale robot for cleaning the near-earth orbit. The study provides an overview of the types and components of the future robots, which at the present time solve the actual problem of the whole world. Our presentation (Click)

Developed a Mobile App with the use of Firebase as a client and material database. Implemented handwritten text recognition using Firebase pre-trained weights and NN model. Programmed App to detect text and provide chemical solutions to the equations. Established Teacher-Student role interface.

Short Description: ChemMy – mobile application is presented as a set of features based on innovative technologies. The project was created in order to make life easier for both teachers and students using the current system, which has many disadvantages, such as not adaptability of the site on different devices and not the constant availability of the site, where grades are set and presentations are laid out. Many students and teachers do not want to spend time searching for a device to access the site or wait until the site resumes. The project enables users to access materials through the convenient menu of the mobile application, as well as to use innovative teaching methods such as teaching through augmented reality, text recognition technologies, which allow users to visualize and analyze data in seconds.

Developed a Mobile App to show interactive 3D objects on the geography book using Vuforia Engine

Short Description: GeoMy - mobile app whose main goal is to help school students learn geography. By pointing the camera at the map of Kazakhstan, students will be able to see 3D models of sights and significant objects. Through simplified learning in an interactive form, students will better absorb information and learn