Ravit Cohen-Segev

Matrix
SCAR: A Multi-Engine Supply Chain Anomaly Radar Driven by Agentic RAG and Real-Time Automation
Ravit Cohen-Segev

Abstract

The SCAR (Supply Chain Anomaly Radar) system is an operational, data-driven platform developed by the Ministry of Innovation, Science and Technology’s Horizon Line Department in collaboration with the National Emergency Authority. Designed for national resilience, SCAR provides real-time monitoring and predictive early warnings for global disruptions affecting Israel’s most critical strategic supply chains.
Moving beyond theoretical AI prototypes, SCAR implements a Dual-Engine Operational Framework built on a fully automated AWS serverless infrastructure. The first pillar is a Deterministic KPI Alerting System that monitors supply chain health – analyzing price volatility, transport latency, and supply shortages. This engine ingests high-fidelity data from a diverse array of authoritative sources, including UN Comtrade and Israel Customs for trade flows, JODI (Joint Organisations Data Initiative) for global energy balances, and USDA for agricultural production metrics.
The second pillar is an Agentic RAG (Retrieval-Augmented Generation) Workflow powered by Amazon Bedrock. Unlike standard chatbots, SCAR utilizes an Intelligent Agent Router that autonomously orchestrates complex queries. By dynamically selecting the optimal data path whether querying structured SQL tables in Amazon Athena or synthesizing unstructured intelligence from expert reports, news feeds (Trading Economics), and real-time web-scraped data (SERP API) the system provides validated, comprehensive answers to strategic interrogations. Ultimately, the chatbot is architected to exclusively deliver localized strategic advice and actionable mitigation tactics for the Israeli market, translating global macro-volatility into domestic-specific strategic guidance.
By prioritizing a “shipping” mindset over traditional research, SCAR demonstrates how Agentic Orchestration and Multi-Source Data Fusion can transform volatile, high-dimensional global data into a mission-critical radar system for national security.

Bio

Ravit Cohen-Segev is a Data Scientist with five years of industry experience, specializing in the application of rigorous analytical research to complex, large-scale data challenges. Although she holds a PhD in Physiology from the Technion, her professional career has been defined by her ability to bridge the gap between deep academic inquiry and the delivery of high-impact business insights. She excels at translating sophisticated scientific methodologies into scalable solutions that drive measurable results.

Agenda

08:45

Reception & gathering

09:30

Opening remarks by WiDS TLV ambassadors

09:45

Keynote session: Prof. Michal Rosen Zvi

10:15

Keynote session: Hadas Grossmon Ella

10:45

Poster pitches

10:55

Break

11:10

Lightning talks session

12:45

Lunch & poster session

13:30

Roundtable session & poster session

14:20

Roundtable closing

14:30

Talk by Hila Paz

14:50

Talk by Dr. Moran Mizrahi

15:15

Closing remarks

15:30

End