Streaming ML for Kotlin/KMP: online regression, bandits, sketches, calibration, and statistics.
Real-time kennel safety intelligence powered by Confluent Cloud, Apache Flink, and streaming ML.
Streaming fraud detection pipeline processing 1M+ transactions/day with sub-100ms latency and 97.3% AUC-ROC. Kafka + Spark Structured Streaming + XGBoost with a Feast feature store and Evidently AI drift monitoring.
Streaming Evaluation Protocol (Prequential + Resource Footprint + Calibration Drift)
Ultra-low latency (<20ms) polyglot dual-track audio + video streaming pipeline featuring zero-copy POSIX shared memory, C11 AVX2 SIMD DSP & NMS, Rust 2024 PyO3 ML runtime, Go Pion WebRTC gateway, and 60 FPS TypeScript HUD.
To associate your repository with the streaming-ml topic, visit your repo's landing page and select "manage topics."