What we do
GDA was founded in 2012 by Dr. John Harer while he was a Professor of Mathematics at Duke University. GDA grew out of research performed at Duke, funded through a variety of sources including DARPA, AFOSR, NSF, NIH, and DTRA.
We solve complex data analysis problems by merging cutting-edge expertise from academia with intelligible demonstrations and solutions that matter to industry and government. As a small business, our agile approach to contract research helps government agencies and larger companies quickly and efficiently investigate important questions before investing resources in large-scale integration and deployment.
Building strong relationships with our partners is our priority. We develop proof-of-principle methods, and then work closely with collaborators to integrate those methods within real-life, real-time application systems. We work within and adapt to myriad data-gathering systems and are comfortable with a mixture of open-source and proprietary data, software tools, and hardware systems.
Our guiding principle is that clients and partners should be able to use our code both as stand-alone products and as a well-integrated component within their own computational systems. Our employees have decades of experience building computational resources from hardware up, including installing and configuring deep learning frameworks on heterogeneous and distributed resources including GPUs and CPUs. Our standard codebase is developed in a continuous integration environment.
GDA mixes agile software design practices with fundamental research practices to construct compelling demonstrations of research solutions to applied problems.
GDA performs on government research and development contracts, either as a prime or as a subcontractor to various larger companies. Agencies we have worked with or are working with include:
- Defense Advanced Research Projects Agency
- Air Force Research Laboratory
- Department of Energy
- Environmental Protection Agency
- National Geospatial-Intelligence Agency
- National Oceanic and Atmospheric Administration
- National Science Foundation
- National Security Agency
- National Institutes of Health
- Department of Homeland Security
- Office of the Secretary of Defense
- National Reconnaissance Office
Our approaches provide a common language to leverage data from heterogeneous data streams to improve AI models over any single modality.
With widespread impact of machine learning approaches, we enhance existing methodologies with computationally efficient shape features and explainable models.
We provide novel analytics for networks of sensors that operate on the edge, communicating with both each other and a central hub to achieve various mission-critical goals.
We specialize in extracting geometric descriptors from data to augment traditional data analysis with sophisticated concepts from contemporary mathematics.
Our shape analytics, data fusion, and machine learning approaches for UAV, rhizotron, or ‘omics data optimize state-of-the-art agriculture research and crop management.
As IoT networks become increasingly feasible to build, we use our expertise in algorithm design and shape analytics to optimize their design and deployment.
Logistics & scheduling
GDA is pioneering the development of tools for planning, scheduling, and logistics that emphasize modularity and resilience in complicated systems of systems.