Menu
Kevin Larrabee
Kevin Larrabee

Kevin Larrabee

Head of Product
DataShapes AI
Bio

Kevin Larrabee joins DataShapes AI as the Head of Product. He recently transitioned as the Chief Data Officer of a U.S. Army Special Operations Command (USASOC) Research and Development organization. He led the Command’s technical partnerships, enabled capability development, and set strategy towards making data an enterprise asset.
Kevin combined his leadership abilities, strategic vision, and technical prowess throughout his career to solve the nation’s most complex problems. He has innovated across the Army Staff, CENTCOM, and USASOC. He was the first data scientist to deploy in support of the Afghanistan National Army Special Operations Command in 2015, create a CENTCOM wargaming and human dynamic modeling team in 2017, and build a multi-functional data team in 2020.
Under Kevin’s dynamic leadership at USASOC, the data directorate has experienced a remarkable transformation. His entrepreneurial spirit has propelled the team’s growth from a modest 3 to a robust 30 personnel while catapulting the data budget from $500K to an impressive $10M. He has overseen eight concurrent machine learning and software development projects, facilitating advanced technology’s rapid adoption, building, and integration.
Kevin holds a B.S. in Business Management from the U.S. Military Academy, an M.S. in Operations Research from the U.S. Naval Postgraduate School, and an MBA from Winston-Salem State University. He travels, invests, and coaches his children’s sports teams in his free time.


See the Unseen: Advancing RF Analytics for Rapid Decision-Making

The increasing proliferation of RF-based technologies across domains such as telecommunications, IoT, and defense has increased the demand for robust analytical methods to extract actionable insights from RF signals. Applying an artificial intelligence approach to detecting, identifying, and learning at the edge enables the characterization of spectral patterns, anomaly detection, and predictive modeling of signal behavior. This presentation explores state-of-the-art methodologies for processing RF time series data, including feature extraction, metadata RF event records, and advanced modeling approaches.
A key challenge in RF analytics is the high dimensionality, volume, and the necessity of domain expertise. These characteristics detract from most data scientists’ ability to effectively analyze and interpret IQ data, as traditional machine learning techniques often struggle with RF signals’ complexity. This discussion examines the real-world need so many industries have to see and manage the spectrum, from defense to critical national infrastructure and sports stadiums. Spectrum monitoring, interference detection, and signal classification in congested environments are all use cases that struggle to extract insights at the edge for near-real-time decision-making. Experimental results demonstrate how modern time series approaches enhance signal interpretation, improve detection accuracy, and support real-time decision-making in RF systems.
This work integrates domain knowledge with data-driven methodologies to highlight the potential of advanced time series analytics to transform RF signal processing. We conclude with a discussion of future research directions, including generative AI and the role of edge computing in real-time signal analysis. This discussion aims to bridge the gap between classical signal processing and modern business applications, offering a roadmap for practitioners and researchers in RF analytics.

Back To Top

Institute for Operations Research and the Management Sciences

5521 Research Park Drive, Suite 200
Catonsville, MD 21228 USA

phone 1 443-757-3500

phone 2 800-4INFORMS (800-446-3676)

fax 443-757-3515

email

Get the Latest Updates

Copyright INFORMS. All Rights Reserved