Declarative Python framework for defining, computing, and alerting on KPIs from SQL or DataFrames — with built-in period-over-period comparisons and automatic trend analysis.
Streaming-native concept drift detection library with four canonical algorithms, drift type classification, retraining window recommendation, and multi-channel alerting — zero mandatory dependencies.
Automatically infers optimal star and snowflake schemas from raw DataFrames and generates production-ready DDL for Snowflake, Redshift, BigQuery, and PostgreSQL in seconds.
Adaptive data stream chunker that dynamically sizes batches based on memory pressure and processing latency for real-time ETL pipelines — with CPU-aware parallel processing and multi-threading.
Python package for real-time hand tracking and gesture recognition using MediaPipe — published on PyPI with a simple API for detecting hand landmarks, counting fingers, and computing inter-landmark distances.
