Turning signals into intelligence
A 24-hour experiment in sensor fusion.
Inside the Hackathon
Self-driving cars don't rely on a single sensor. Neither do the most advanced AI systems being built today.
Whether it's combining cameras with LiDAR, weather forecasts with live telemetry, or video with structured data, the next generation of AI is being built by bringing multiple sources of information together to make better decisions.
That approach is called sensor fusion: the focus for this year's Hackathon.
Over the course of 24 hours, cross-functional teams from across the U.S. and the Balkans came together to design, build, and pitch working AI prototypes centered around one idea: create an insight that couldn't exist from a single data source alone. Teams were challenged to build multi-modal AI solutions that combined different forms of data: video, images, telemetry, structured data, audio, and text to solve real business problems in entirely new ways.
Twenty-four hours later, four working prototypes had emerged. Each tackled a different industry, a different business challenge, and a different application of AI. Together, they offered four distinct perspectives on what sensor fusion can make possible.
Four teams. Four industries. One challenge.
FormIQ
How do you coach physical movement without an expert standing beside you?
FormIQ combines computer vision, biomechanical reference data, and large language models to analyze movement and generate personalized coaching recommendations. By measuring 33 skeletal landmarks against professional benchmarks, users receive immediate feedback on where movement breaks down and how to improve it.
While the prototype demonstrates value in sports performance, the same approach could extend to workforce training, manufacturing safety, physical therapy, and employee onboarding, where consistent coaching at scale has traditionally been difficult and expensive.
FieldSense
A field shows 19% soil moisture. Rain is forecast for Friday. Do you irrigate now or wait?
Every growing season requires thousands of decisions, many of which happen in a matter of hours.
FieldSense combines soil sensor data, weather forecasts, and crop intelligence to answer questions like these with greater confidence, delivering recommendations around irrigation, nutrient management, and crop health.
CropTop
Farmers generate enormous amounts of operational data, yet much of it never influences financing or insurance decisions.
CropTop combines farm operations, sustainability metrics, and agricultural data into an AI-powered agricultural credit score, helping lenders, insurers and ecosystem partners better understand farm-level risk while giving producers clearer visibility into the decisions that influence their financial future.
Kairos
When disaster strikes, every hour matters.
Kairos combines weather forecasts, historical storm paths, terrain data, infrastructure maps, and live telemetry to recommend where disaster recovery organizations should stage equipment close enough to mobilize quickly, while remaining outside the projected impact zone.
The result was a smarter approach to protecting assets without sacrificing response time.
Why sensor fusion, why now?
Every industry has data. The challenge is understanding which signals matter, how they relate to one another, and what they should tell us.
That was the thread running through every project at this year's Hackathon.
A weather forecast can predict a storm, but it can't recommend where recovery equipment should be staged. A soil sensor can measure moisture, but it can't explain how upcoming weather will affect crop health. A video can capture movement, but it can't determine whether someone's technique is correct without something meaningful to compare it against.
Sensor fusion isn't a new idea. What's changed is our ability to apply it.
~Jason Harris, Principal
That shift opens the door to new kinds of products, new ways of solving problems, and ultimately, new opportunities for organizations.
Winner: FormIQ
FormIQ earned this year's top honor because it demonstrated the promise of sensor fusion in a way that was both technically impressive and immediately practical.
The prototype combined sophisticated AI techniques with a clear real-world application, showing how a single approach could create value across multiple industries.
Looking Ahead
This year's Hackathon reinforced something we believe will shape the next generation of AI: the breakthrough often isn't a new model. It's the ability to connect information in ways that create entirely new insights.