Pulse is a tiny real-time data streaming framework (mini Flink/Beam) in Rust. Async (Tokio), pluggable operators, state, and I/O. Fast, modular, local-first.
-
Updated
Oct 29, 2025 - Rust
Pulse is a tiny real-time data streaming framework (mini Flink/Beam) in Rust. Async (Tokio), pluggable operators, state, and I/O. Fast, modular, local-first.
Embedded-first, Arrow-native stream processing engine in modern C++ (C++23). Run SQL streaming pipelines in-process in milliseconds, or submit the same file to a distributed cluster: exactly-once, event time, state as open Arrow data, deterministic replay.
Delayed conversion labels: freshness, correctness, corrections, processing lag, and retained state.
Real-time decisions on an electricity distribution network. Deterministic code owns whether a window is closed; models only forecast, score and rank. Seven claims proved offline and against a deployed AWS estate — 0 rows published early of 3,779, 41 gate mutations refused, all six erasure legs confirmed.
Evidence-bound ad click aggregation reference: event-time windows, fenced workers, immutable publications, explicit corrections, and PostgreSQL authority.
Kafka -> Spark Structured Streaming -> Delta, with event-time windowing and watermarking made runnable: a dependency-free reference implementation you can execute without a JVM
Kafka-compatible streaming demo for real-time birding intelligence: event-time windows, late-event handling, stateful deduplication, hotspot metrics, and target-species alerts.
Interactive F# lab for watermark lag, late events, idle partitions, incomplete windows, and delayed alerts.
Real-time anomaly detection on payment streams with PyFlink: event-time windows, keyed state, watermarks
Demonstrates architectural pattern implementations in Apache Beam (Java SDK) to restore event-time ordering on out-of-order keyed streams.
PyFlink streaming pipeline: Kafka → 5-minute tumbling event-time windows → per-customer spend aggregation. Handles late-arriving events with watermarks and BoundedOutOfOrderness strategy.
Proyecto integrador de Data Streaming: detección de velocidad transaccional con Kafka y Apache Beam
Cross-engine output validation for event-time streaming pipelines.
BTCUSDT top-of-book microstructure signal study for short-horizon event-time midprice direction prediction.
Deterministic event-time watermarks with terminal precedence and atomic storage providers.
Build deterministic, boundary-aware cell and edge rollups from trajectory observations.
Deterministic alignment and watermarking for multimodal event streams
To associate your repository with the event-time topic, visit your repo's landing page and select "manage topics."