# Stanislav — DATA SCIENTIST

> Building services from data to deployment

## Projects

- [Best Free LLM](https://bylupiner.com/en/free-llm) — Daily check of which LLMs are free, ranked against paid.
- [Pixel Choice](https://bylupiner.com/en/pixel-choice) — Artwork where 1,024 cells each choose one of 24 shades.
- [RAG Movie Recommender](https://huggingface.co/spaces/lupinerh/movie-recommender-app) — Movie picks from a text query: SVD, ChromaDB, RAG.
- [Hybrid Sentiment Chatbot](https://huggingface.co/spaces/lupinerh/sentiment-chat-app) — Sentiment and per-word emotion with ML, BERT, LLM.
- [Educational AI Assistant Bot](https://t.me/prepgenius_bot) — Telegram bot: ML mock interviews, code grades, Anki.
- LTV & Churn Predictive Pipeline — CatBoost LTV and churn forecasts on Airflow.
- E2E Support Ticket Automation Bot — BERT sorts support tickets and picks template replies.
- Fraud Detection in Gaming Telemetry — Flags cheating players in gameplay telemetry.
- Weekly Revenue Forecasting & Pricing Simulator — Weekly revenue forecasts plus a pricing simulator.

## Experience

### Independent Developer — Data Scientist & ML Developer (2023 – Present)
- RAG Movie Recommender (SVD + ChromaDB + RAG): movie suggestions from text, live on HuggingFace
- Sentiment Chatbot (BERT + LLM): per-word emotion highlighting, live on HuggingFace
- Interview Bot (Telegram, Anki): DS/ML mock interviews with spaced repetition
- Anki AI Assistant: auto-grades answers, generates tests and audio podcasts

### Plarium — Data Scientist (05/2019 – 09/2022)
- LTV Pipeline (Airflow): daily forecasts for marketing budget planning
- Churn Model: VIP segment churn prediction
- Revenue Forecasting: weekly revenue forecasts + pricing simulator
- Fraud Detection (CatBoost): found new cheating patterns in gameplay telemetry
- Support Bot (FastAPI, Docker, MongoDB): outperformed built-in search, reduced first-line load
- Auto-Tagging: optimized Transformer for ticket auto-tagging
- Optimal Push: personalized notification timing by behavioral clusters
- Data Monitor + Tableau: automated anomaly detection

## Skills

- **DS & ML:** Python, SQL, PySpark, Pandas / NumPy, Scikit-learn, CatBoost, SHAP / LIME, Tableau
- **NLP & LLM:** LLM / RAG, Transformers (HuggingFace), BERT, Vector Databases (ChromaDB / Faiss)
- **MLOps & Infrastructure:** Docker, Apache Airflow, REST API (FastAPI / Flask), PostgreSQL / MongoDB, Hadoop / Git

## Contact

- Email: lupinerh@gmail.com
- Telegram: @lupinerh (https://t.me/lupinerh)
- GitHub: https://github.com/lupinerh
- Hugging Face: https://huggingface.co/lupinerh
- Full CV: https://bylupiner.com/resume.md | https://bylupiner.com/Stanislav_Data_Scientist.md
