I will develop a specialized software with live feeds concurrent threads

WPF desktop app integrating the Rithmic RAPI SDK for live data ingestion across multiple simultaneous streams, with thread-safe processing that keeps the live-feed thread never blocking on downstream work. Implemented configurable anomaly-detection logic (unusually large or clustered events, relative to a rolling baseline) with tiered alerting and rolling-window trend tracking the same pattern used for fraud alerts, equipment-monitoring dashboards, or any system watching for "activity that deviates from normal." A a two-stage filter (a statistical scaling stage plus a validity/quality-gate stage) for a larger analytics pipeline, correctly identifying which stage each rule belonged to and implementing the piece that had never actually been built. The underlying pattern narrow a large dataset down to the subset that's both statistically relevant and reliable enough to trust generalizes to recommendation engines, quality-control pipelines, and any system that has to separate signal from noise at scale.