Machine Learning Research · Data Science · Software Engineering
Software Engineering undergraduate at HSE University · Data Analyst at Yandex

🧪 Research:  Probabilistic ML  Language Modeling  Trustworthy AI
💼 Industry:  Product Analytics  Experimentation  Data Systems

Email

I work at the intersection of machine learning research, data science, and software engineering. My research interests include statistical and probabilistic learning, language modeling, information retrieval, and the reliable evaluation of AI systems. My projects combine method implementation, controlled experiments, statistical analysis, and reproducible artifacts.

Alongside research, I work as a Data Analyst at Yandex on product analytics, experimentation, predictive modeling, and production data pipelines. This industry experience reinforces my interest in methods that are not only mathematically grounded, but also carefully evaluated and reliable in practice.

Research completed at the CUHK-Shenzhen Elite Undergraduate Summer Camp 2026 under the supervision of Associate Professor Shenghao Yang.

• Implemented Context-Tree Weighting and practical text context-tree models from scratch.
• Compared finite-memory models with GPT-2 on WikiText-2 and investigated whether the performance gap comes from limited context or sparse statistics.
• Reached 2.162 bits per character with the context-tree model, versus 1.154 BPC for GPT-2 small; the depth-7 conditional-entropy estimate was 1.205 BPC, pointing to data sparsity as the central limitation.
• Packaged the work as reproducible research code with tests, experiment artifacts, a full report, slides, and citation metadata.

Research report · Presentation · Reproduction guide

An evidence-grounded pipeline for assessing corporate sustainability reports and public documents across 15 ESG indicators.

• Built hybrid retrieval with BM25, vector search, and reciprocal rank fusion.
• Explored direct prompting, RAG variants, Writer-Critic workflows, ensemble methods, multi-agent orchestration, and LoRA fine-tuning.
• Evaluated scores and supporting evidence with bootstrap resampling, MAE, citation similarity, evidence-type accuracy, and McNemar's test.
• Research-course project at HSE University; manuscript in preparation.

A full-stack collaborative playlist application with competitive voting and real-time room synchronization.

• iOS: Swift, UIKit, Spotify API, SVIP architecture
• Backend: Go, REST API, WebSockets
• Data and infrastructure: PostgreSQL, Docker Compose

A curated collection of six substantial projects from the Machine Learning, Deep Learning, and Applied Data Analysis tracks at HSE University.

• Benchmarked XGBoost, CatBoost, and LightGBM and analyzed clustering failure modes on tabular data.
• Built classical NLP pipelines with sparse text features and controlled preprocessing experiments.
• Studied neural-network optimization and regularization in PyTorch, reaching 8.676 test RMSE in a regression task.
• Implemented face verification with pair classification and triplet loss, reaching 0.8682 ROC-AUC.
• Compared waveform, recurrent, and spectrogram-based audio classifiers, reaching 0.9401 validation accuracy.
• Implemented cWGAN, conditional VAE, diffusion, and RealNVP models for synthetic scientific data.

• Bachelor's student in Software Engineering at HSE University
• Data Analyst at Yandex, working across experimentation, predictive modeling, analytics engineering, and AI-assisted analytical workflows
• Former Python programming instructor for a class of 25 students
• Winner, Rosatom Mathematics Olympiad

Machine learning and AI: Python · PyTorch · scikit-learn · probabilistic modeling · transformers and attention · CNNs and transfer learning · generative models (VAE, GAN, diffusion) · NLP and information retrieval · computer vision

LLM systems and evaluation: RAG · LoRA fine-tuning · multi-agent systems · AI-agent workflows · evidence-grounded generation · LLM evaluation · bootstrap analysis and statistical testing

Experimentation and product analytics: A/B testing · experiment design · statistical validation · feature engineering · predictive and relevance scoring · anomaly detection · metric hierarchies · event instrumentation · customer journey mapping · BI dashboards

Data engineering and systems: SQL · Pandas · ETL and data pipelines · distributed computing · PostgreSQL · Docker · Go · REST APIs · WebSockets

Software engineering: Go · C++ · C# · Java · Assembly · Swift/UIKit · API design · testing · reproducible tooling

Languages: Russian (native) · English (C1) · German (B1)