cv

AI for map Geolocation and Extraction to find Critical Minerals (AIM)

sponsor: DARPA CriticalMAAS
advisor: Dr. Yao-Yi Chiang

• Conducting research aimed at segmenting map content and identifying legend areas to improve model performance across diverse map styles
• Analyzing and labeling maps, focusing on those with suboptimal styles for better legend representation
• Developing models to detect polygon legend items and advancing techniques for handling complex legend layouts to improve fine legend detection accuracy

HAYSTAC: Context-Aware Trajectory Anomaly Detection

sponsor: IARPA
advisor: Dr. Yao-Yi Chiang

• Contributed to the project aimed at establishing models of "normal" human movement across time, locations, and people to detect anomalous activities
• Assisted with the design and illustration of visual aids for the paper, supporting the research in anomaly detection models
• Participated in the refinement of a context-aware approach that incorporates elements such as subject identity and nearby points-of-interest (POI) for better trajectory anomaly detetion

Methods for Detecting Critical Events in System Log Data

advisors: Oleg Gorokhov, Mikhail Petrovsky

• Analyzed the core methods of event logs processing based on NLP techniques
• Developed preprocessing method for the raw log data based on natural texts preprocessing and detecting semantic information in logs
• Designed a CNN for detecting anomalies in clusters of log data
• Proposed a new approach of effective clustering for huge log files based on extraction sequences of events with process mining algorithms
• Achieved 98% accuracy on finding anomalies in preprocessed data and 82% within rawdata

Evaluation and Comparison of Algorithms for the Connect6 Game with Monte Carlo Method

advisor: Olga Oparina

• Conducted a benchmark analysis between several classical heuristic algorithms which are used for evaluating stochastic processes
• Measured accuracy on considered methods and state Monte Carlo method as the best for providing solitions for Connect6 game

Ph.D.

University of Minnesota

Computer Science

• Key words of a future PhD thesis: spatiotemporal data prediction & forecasting, natural language processing, computer vision

B.Sc.

Lomonosov Moscow State University

Computer Science

• Thesis: Process Analytics Methods for Anomaly Detection Tasks in System Log Data

Context-Aware Trajectory Anomaly Detection

The 32nd ACM International Conference on Advances in Geographic Information Systems (SIGSPATIAL'24)

Evaluation and Comparison of Algorithms for the Connect6 Game with Monte Carlo Method

Journal of the LXX Open international Student's Scientific Fair of Moscow Polytechnic University

Machine Learning Researcher

Rotec Digital Solutions, Moscow, Russia

Developing ML models for detecting anomalies in the operation of equipment, solving predictive maintenance problems and conducting research

• Frameworks: PyTorch, scikit-learn, PostgreSQL, ClickHouse
• Platforms: MLflow, DVC, Airflow, Kedro, Prometheus, Cookiecutter, Grafana, Optuna
• Technical skills: Docker, Git, command line, Anaconda

Full-stack Java Developer

MOLNET, Moscow, Russia

Designing web systems from scratch providing complete production cycle from servlets to interfaces

• Development: Java, Spring Framework, JavaServer Pages (JSP), Maven, RESTful services/APIs, PostgreSQL
• User interfaces: React, Bootstrap, TypeScript
• Technical skills Docker, Git, command line

Freelance Writer

JetBrains Academy, remote

An author of maths and algorithmic topics for computer science courses

• Areas: probability theory, statistics, data structures, algorithms

TV NeuroTech hackathon. 1st prize

Case: to find, formalize and explore the statistical validity of the dependencies of indicators of the quality of life of the population (income, demography, credit burden, etc.) on the indicators of state resource provision | Data: official statistics of the Russian Federation for the last 10-15 years | My task was to detect and clean outliers in the data.