June 19, 2025
Elbit America engineers Brandon Kessler and Josh Haley are helping shape the future of artificial intelligence (AI) within the company — and people are noticing.
The duo recently earned recognition at the SPIE Defense + Commercial Sensing Conference in Orlando, Florida for their novel approach to AI performance modeling, which they shared during the event and received the Best Poster Award — honoring both the innovation itself and the clarity with which it was presented.
Pushing the boundaries of AI performance
Kessler and Haley's project explored using photorealistic synthetic data to train deep learning networks for object detection and recognition. They also developed a performance model that enables self-improvement and stronger simulation and trust capabilities.
“We’re excited to contribute to a growing body of work that improves how AI models are trained, evaluated, and trusted,” said Brandon Kessler, Principal Algorithm and AI Engineer at the company.
Their findings were detailed in a paper, “Contextual Performance Metrics,” published in the conference proceedings.
“Being able to model where AI might fail — and then teach it to get better — is a big step forward in making these systems more reliable and mission-ready,” said Josh Haley, Elbit America Staff Algorithm and AI Engineer.
Why it matters
This recent work by Kessler and Haley not only advances Elbit America’s technical leadership in the AI space, but it also strengthens the trust our customers can place in our products that incorporate AI. Elbit America understands the unique constraints the end-users of our solutions face during military operations.
Matt Steenman, Elbit America’s Chief Technology Officer, explained that while most commercial AI solutions rely on vast data sets to train their algorithms, as well as massive amounts of cloud computing resources, Elbit America engineers are designing AI solutions that can operate with limited communications bandwidth, and cloud computing resources. Despite these constraints, their work is also optimized for use in contested environments.
Steenman said, “I am incredibly proud of the effort that our teams our doing in the area of AI, and this publicly released work is just a glimpse into the advanced AI and methodologies we are working on.”
Interesting in joining Elbit America's team of engineers? View open positions on our Careers page here.