Hi, I’m Timur 👋

I’m a Master’s student in Artificial Intelligence at Chung-Ang University, Seoul, researching how to make machine learning more efficient and intelligent. My current work focuses on feature selection and unsupervised learning, and I am deeply interested in Large Language Models and their Optimization.

Before diving into AI research full-time, I spent a year as a C++ Software Developer at YADRO, optimizing LTE physical layer software for 4G networks. That experience taught me the value of efficient algorithms and real-time systems - lessons I now apply to machine learning.

🔬 What I’m Working On

Master’s Thesis: “Maximizing Joint Entropy and Pattern Discrimination for Unsupervised Feature Selection” Developing automated methods to identify the most informative features in high-dimensional data without labels.

Recent Project: Optimizing Vision-Language Models for Production Benchmarking LLaVA, Qwen-VL, and PaliGemma to find the sweet spot between performance and efficiency using quantization and pruning. Spoiler: 4-bit quantization is a game changer!

💡 Research Interests

Current Focus:

• Large Language Models (LLMs) & Optimization
• Efficient learning in resource-constrained environments
• Decision-making algorithms

Research Interests:

• Unsupervised feature selection & information theory
• Data imputation & handling missing values
• Efficient neural network architectures
• Computer vision & generative AI

📚 Recent Publications

ChainImputer: A Neural Network-Based Iterative Imputation Method Using Cumulative Features MDPI Symmetry, 2025 | Paper

Diffusion Model-Based Generative Pipeline for Children Song Video IEEE ICCE, 2025 | Paper | Code

Object Detection Review for Autonomous Vehicles IEIE, 2024 | Paper

🎓 Background

M.S. in Artificial Intelligence - Chung-Ang University (2024-Present) CAYSS Scholarship recipient | GPA: 4.22/4.5 AutoML Lab

B.S. in Computer Science - Electrotechnical University “LETI” (2018-2023) GPA: 4.3/5.0 | Exchange semester at Inha University Helped launch Russia’s largest Robotarium

C++ Software Developer - YADRO (2023-2024) Optimized LTE physical layer software using AVX instructions Specialized in real-time telecommunications systems

💻 Technical Stack

Languages: Python, C/C++, MATLAB, Dart, SQL AI/ML: PyTorch, TensorFlow, Scikit-learn, OpenCV, Hugging Face Transformers Tools: Git, Linux, Docker, W&B, Flutter, Qt Specializations: Deep Learning, LLM Optimization, Feature Engineering, Computer Vision

🏆 Highlights

• Yandex Algorithms Bootcamp 3.0 (2023): Top 200/2000 participants
• CAYSS Scholarship at Chung-Ang University (2024-Present)
• Led 3-developer team on enterprise telecom project at YADRO
• Teaching Assistant for Computer Vision course (Spring 2025)

🚀 Featured Projects

VLM Optimization - Comprehensive benchmark of quantization and pruning strategies for Vision-Language Models.

Feature Selection Library - Research implementations supporting multiple publications

Menu Management App - Production Flutter/Firebase application with real-time sync

L1-L2 Communication Visualizer - Enterprise telecommunications analysis tool (YADRO proprietary)

📫 Let’s Connect

I’m always interested in collaborating on research related to:

• LLM optimization and efficiency
• Efficient machine learning methods
• Real-world AI applications

Email: timurkhairulov@cau.ac.kr GitHub: KhrTim LinkedIn: timur-khairulov Google Scholar: Publications