Analytics, machine learning and process automation. MSc student in Artificial Intelligence at HSE University, Faculty of Computer Science. I work with industrial data: equipment telemetry, failure prediction and turning routine analytics into tools.
industrial-telemetry-anomaly
Anomaly detection in turbofan telemetry: five methods compared on NASA C-MAPSS. Isolation Forest warns about all 100 engines 12 cycles before failure with one false alarm per thousand healthy readings.
Python, DuckDB, scikit-learn, Streamlit
Slovo
A native macOS app that transcribes meetings and lectures and writes minutes entirely on-device, with no internet and no sign-up.
Swift, SwiftUI, WhisperKit, llama.cpp
More this autumn: a remaining-useful-life prediction service, open-registry ETL pipelines and company financial risk scoring.
Python, Polars, pandas, DuckDB, SQL, scikit-learn, LightGBM, PyTorch, Streamlit, Swift
Telegram @iamartempn