Mr. Ulman led the development and operational deployment of numerous software systems, tactical decision-aids, and data analysis tools utilizing Bayesian data fusion, Bellman decision algorithms, and machine learning models for classification, clustering, and anomaly detection.
Mr. Ulman joined Metron in 2005, helping maintain Metron’s first operational deployment of its Nodestar multiple target tracking software in submarine combat systems. In 2008, Mr. Ulman developed a highly parallel real-time data processing engine for maritime shipping kinematic anomaly detection. The system ingested global position data on tens of thousands of vessels and analyzed the data for anomalous shipping patterns.
In 2012 Mr. Ulman released Glimpse, Metron’s open-source high-performance data visualization library. Glimpse provides Java and OpenGL software tools for rapidly building applications that utilize GPU hardware to enable interactive, real-time filtering, scaling, and shading of large, multi-dimensional datasets. Mr. Ulman also maintains Web Glimpse, which utilizes WebGL to deliver a subset of Glimpse’s high performance data visualizations for web applications. These software libraries have been utilized in a wide variety of high-performance data analysis applications in domains including multi-dimensional radar and sonar, UUV mission planning, air traffic control, nautical charting, and maritime domain awareness.
Glimpse was a key component in software tools Mr. Ulman developed in 2013 to enable novel narrowband sonar gram search techniques to reduce the manpower necessary for ASW search. Mr. Ulman also combined Glimpse data visualization technology with Bayesian particle filters for analysis of sensor data from unmanned underwater vehicle platforms.
While completing his Master of Science in Computational Science at George Mason University in 2013, Mr. Ulman applied CUDA to accelerate particle filter tracking applications using GPU hardware, perform M-of-N range rate filtering to detect weak targets in real-time radar data, and train convolutional neural network classifiers on flight path and weather data.
In 2016, Mr. Ulman led a Metron software team working with Dr. Siver and Mr. Gurley to develop prototype mission planning and decision support software utilizing Bellman decision algorithms. Mr. Ulman now manages the software teams operationally deploying these capabilities for multiple government sponsors. The scale of these efforts required Mr. Ulman to build robust DevOps and CI/CD infrastructure. In 2022, the procedures and software tools developed to support these efforts were developed into Bell Tower, a set of software tools and libraries to help Metron teams deploy computationally intensive web applications backed by robust DevOps processes.