A. Personal Statement

My career trajectory reflects a sustained commitment to applying advanced information technology and machine learning methodologies to complex real-world problems, progressing from software engineering and large-scale systems development to scientific research and leadership in applied data science. My professional background combines practical experience in designing scalable data systems with a growing focus on methodological research in artificial intelligence and machine learning.

I am the co-founder of the National Health Data Science Lab, a scientific initiative focused on the application of data science and machine learning approaches to healthcare and biomedical research. In this capacity, I contribute to research direction and coordinate development of scalable and reproducible computational frameworks.

My academic training includes a Master’s Degree in Information Technology with a specialization in Software Engineering, providing a strong foundation in algorithm design, data processing, and scalable software architectures. Throughout my career, I have developed and implemented large-scale data processing pipelines and machine learning workflows, emphasizing reproducibility, modular design, and robustness — principles that directly support modern computational research practices.

I am an active member of leading professional organizations, including Senior Membership in the Institute of Electrical and Electronics Engineers (IEEE), professional membership in the Association for Computing Machinery (ACM), and membership in the Association for the Advancement of Artificial Intelligence (AAAI). These affiliations reflect my continued engagement with the broader scientific and engineering communities.

Since 2024, my research activities have focused on artificial intelligence, machine learning, and data science, with an emphasis on predictive modeling, model interpretability, and methodological aspects of machine learning systems.  My peer-reviewed publications demonstrate the application of advanced AI and ML techniques to complex data environments and highlight my ability to design and evaluate computational methodologies in interdisciplinary settings.

My current research direction focuses on methodological challenges associated with applying machine learning to complex biomedical and genomic datasets, particularly issues related to reproducibility, robustness, and uncertainty-aware evaluation. The goal of my work is to develop scalable and reliable ensemble machine learning frameworks capable of improving the stability and interpretability of AI-driven biomedical analysis, including genomic risk modeling for neurodegenerative diseases such as Alzheimer’s disease.

My combined expertise in software engineering, machine learning research, and organizational leadership positions me to successfully execute computationally intensive research projects and contribute methodological advances at the intersection of artificial intelligence and biomedical data science.

B. Positions, Scientific Appointments, and Honors

Positions and Scientific Appointments

2024 - Present   Research Lead, National Health Data Science Lab

2022 - 2024   Graduate Researcher, Kharkiv National University of Radio Electronics

С. Research Contribution and Accomplishments 

1. My research contributions focus on the application of artificial intelligence and machine learning methods to predictive modeling in complex and heterogeneous data environments. My early work concentrated on methodological aspects of model development, data preprocessing, and evaluation in settings characterized by high variability and uncertainty, which later motivated my current interest in improving robustness and reproducibility in machine learning systems.

In my publication, “Bitcoin Price Prediction Using the Boosting Algorithm,” I investigated the effectiveness of machine learning approaches for modeling nonlinear dependencies in financial time series data. The study involved comprehensive data preprocessing, feature engineering, model training, and hyperparameter optimization. The results demonstrated that properly configured machine learning models can maintain stable predictive performance despite significant market volatility. This work contributed to understanding how model configuration and feature representation influence predictive stability in complex data environments.

a.  Anton Naumov; Afanasieva Irina, Onyshchenko Konstantin; “Bitcoin Price Prediction Using the Boosting Algorithm”; The 2nd International scientific and practical conference “Science and society: modern trends in a changing world” (January 22-24, 2024) MDPC Publishing, Vienna, Austria. 2024. ISBN: 978-3-954754-01-4 https://sci-conf.com.ua/wp-content/uploads/2024/01/SCIENCE-AND-SOCIETY.-MODERN-TRENDS-IN-A-CHANGING-WORLD-22-24.01.24.pdf#page=197

2. In my second publication, “A Study of the Effectiveness of Using Information Technology Based on Artificial Intelligence for Bitcoin Price Forecasting,” I extended this work by incorporating natural language processing and sentiment analysis into predictive modeling. Using tools such as TextBlob, VADER, Flair, spaCy, and BERT in combination with LSTM neural networks, I evaluated the impact of sentiment-derived features on prediction performance. The findings demonstrated that integrating heterogeneous data sources can improve predictive accuracy and highlighted the importance of multimodal data integration in machine learning applications.

a.  Anton Naumov; Victoria Vysotska; Kirill Smelyakov; Valentina Shtanko; “A Study of the Effectiveness of Using Information Technology Based on  Artificial Intelligence for Bitcoin Price Forecasting”; 2024 IEEE 19th International Conference on Computer Science and Information Technologies (CSIT); 2024; DOI: 10.1109/csit65290.2024.10982646  

3. My subsequent research addressed the challenge of interpretability in complex machine learning systems. In the work “Improving Model Explainability in Dynamic Facial Expression Recognition for Hybrid Intellectual Systems,” I contributed to the development of explainable AI methodologies aimed at improving transparency in deep learning models operating under real-world conditions. The study explored graph-based approaches and attention mechanisms and applied explainable AI techniques such as Attention Attribution, Feature Ablation, Grad-CAM, and Attention Rollout to improve model interpretability without compromising predictive performance. 

a. Vysotska, V.; Smelyakov, K.; Chupryna, A.; Kochkina, A.; Pliekhova, G.; Naumov, A. "Improving model explainability in dynamic facial expression recognition for hybrid intellectual systems"; Computational Linguistics Workshop at CoLInS 2025; 2025; DOI: 10.31110/colins/2025-1/009

Collectively, these studies established my research interest in methodological aspects of machine learning, including data representation, model evaluation, and interpretability in complex data environments. These experiences directly inform my current research direction focused on improving robustness, reproducibility, and methodological reliability of computational approaches applied to complex biomedical and genomic datasets.