This article explores the theory behind explainable car pricing using value decomposition, showing how machine learning models can break a predicted price into intuitive components such as brand premium, age depreciation, mileage influence, condition effects, and transmission or fuel-type adjustments.

Двухэтапная рекомендательная система на MovieLens с отдельным отбором кандидатов, переранжированием и API выдачи.

Сервис геокодирования московских адресов с нормализацией, сопоставлением, ранжированием кандидатов и Docker-запуском.

Система поиска похожих автомобилей по тексту, метаданным и визуальным признакам с оценкой качества и FastAPI API.

RAG-контур с переформулировкой запроса, переранжированием, цитированием и проверкой ответа по найденным источникам.

Стенд для сравнения поиска BM25, векторного поиска, объединения сигналов и переранжирования на FAQ-корпусе.

PersonaTTS is a personalized neural text-to-speech system that learns a user’s vocal persona from a short speech sample and generates natural speech for arbitrary input text.

Система ответов на финансовые вопросы по документам с поиском, переранжированием и генерацией по найденным источникам.

Data-driven modelling framework for utility-scale solar PV inverters, covering digital twins, forecasting, anomaly detection, and maintenance analytics.

End-to-end applied ML system for predicting next-cycle manufacturing cycle times. Includes Spark-based ETL, leakage-safe temporal splits, baseline regressors, LSTM sequence modeling, and SHAP-based explainability under real operational constraints.

Production-style ML model monitoring system with drift detection, delayed labels, retraining and safe promotion.

Fintech applied ML MVP built with OpenClaw: smart collections prioritization with strategy recommendations (3-min demo available)

B2B SaaS applied ML MVP built with OpenClaw: churn risk scoring + next-best-action copilot (3-min demo available)

Adult Income Drift Lab conducts a comprehensive model stability analysis under demographic covariate shift, combining statistical drift detection with performance and calibration evaluation on real-world census data.

Schema-driven decision/control API for AI/ML-related actions, with deterministic allow/block outcomes, explicit validation, and structured audit logging.

Graduate ML and data processing projects covering time series analysis, recommender systems, logistic regression, and feature selection.