Speaker
Description
The EIC offers an opportunity to rethink scientific workflows by combining streaming readout, artificial intelligence, and distributed computing into closed-loop workflows that connect data taking, processing, physics analysis, and operational feedback. We present SPINDLE (Streaming Physics INtelligence and Data Lifecycle Engine), an effort to develop AI-enabled workflows that integrate streaming detector data, FAIR data lifecycle management, anomaly detection, and root-cause analysis. Building on successful demonstrations of real-time data transfer and processing between Brookhaven National Laboratory and Jefferson Lab, SPINDLE aims to combine machine-learning methods with detector knowledge graphs to identify changes in detector performance, uncover their underlying causes, and support human-validated operational responses. Using detector systems at Jefferson Lab as testbeds and leveraging capabilities of the upcoming High-Performance Data Facility (HPDF), SPINDLE explores how streaming orchestration and AI-driven detector intelligence can enable automated calibration, validation, and continuous optimization. This work aims to advance integrated, closed-loop scientific workflows for ePIC, accelerating the path from experimental data to physics results while maximizing the scientific reach of the EIC.