Data Scientist | NLP & LLM Evaluation

Master of Information Technology, Whitireia New Zealand (2025). Research focus: readability assessment of content generated by large language models, with particular attention to age-appropriate narrative texts.

Multidisciplinary background: linguistics → marketing → big data → applied IT research.

Based in Moscow.

• Readability of LLM-generated narrative text
• Comparative evaluation of LLMs for children's content generation
• Applied statistical analysis (ANOVA, Tukey HSD) on NLP output
• Classical readability metrics: Flesch–Kincaid, Coleman–Liau, SMOG, Gunning Fog, ARI, Flesch Reading Ease

Master's research (2025). Comparative analysis of narratives produced by 19 large language models against 16 classical fairy tales, evaluated with textstat and statsmodels.

• Story-generation-process — generation pipeline across 19 LLMs and 16 source tales
• LLM_readability_score_analysis — statistical evaluation (ANOVA, Tukey HSD) of readability metrics

Key finding: no single model fully replicates the readability of authentic fairy tales, but individual LLMs excel in specific readability dimensions.

Language: English · Русский