Russian/English/Estonian/Finnish/Swedish phonetic algorithm based on Soundex and Metaphone
Fast Russian Text normalization for TTS using only RegEx.
Python course for 2nd year NLP students at NRU HSE, 2018-2019
NLP Section of the Data Science course, NRU HSE
Python for students in humanities, NRU HSE, 2018-2019
Research question: academic benchmarks (MMLU, MT-Bench, Arena) don't predict how a model behaves in real office work — pressure, gray zones, tough conversations. OfficeIQ Bench measures exactly that. 100 scenarios · 10 criteria · LLM-as-a-judge · fully reproducible runs.
Research sandbox for ablation studies of RAG components (chunking, embeddings, retrieval, LLMs) on Russian-language corpora.
/ru/ConceptNet5.7 Python wrapper
Автоматизирует модерацию резюме на русском языке с помощью LLM. Для модерации используются разные llm модели.
A Claude Code skill that catches Russian-language bugs in code: pluralization, case agreement, terminology drift, transliterated identifiers, language mixing. Complements talkstream/ru-text.
Автоматическое обезличивание персональных данных (ПД) по 152-ФЗ. Готовый MCP-сервер и API для безопасной работы ИИ-агентов, n8n и LLM.
🤖 Build engaging conversations with this intelligent chatbot designed for seamless interactions in English and Russian.
Self-hosted batch LLM pipeline for analyzing customer feedback from Excel. Upload xlsx, describe the task, configure output fields — get structured results. Works with any OpenAI-compatible API and Ollama.
Self-hosted Russian PII anonymization gateway for safer AI workflows.
Превращает архив Telegram-чата в книгу о ваших друзьях. Экспорт → статистика → анализ через Claude → отчёт, после которого все будут спорить, кто там самый «душа компании».
Telegram bot that classifies channel posts with a fine-tuned RuBERT model and sorts them into topics of a forum it creates — plus the labeled dataset and notebooks behind the model.
Russian IT Community Conversational Corpus (2017-2026) · Zero-PII Curation Platform, Multi-turn SFT, DPO, RAG Knowledge Base, Streamlit Studio & RTX 3060 LoRA benchmark
Neural morpheme segmentation for Russian. Lightweight, CPU/ONNX, C++ core.
Prompting vs RAG vs LoRA for ICD-10 coding of Russian clinical complaints (RuMedTop3), with honest baselines and costs
Research-backed Russian AI humanizer skill for Claude Code. Blind test vs humanizer, humanizer-ru, slopotron: fewest meaning errors (0.03/text). Rules from 1851 human vs 4189 AI texts (Claude, GigaChat, YandexGPT). Хьюманайзер для русского текста: убирает признаки нейросети, не теряя смысл.
To associate your repository with the russian-nlp topic, visit your repo's landing page and select "manage topics."