Pioneering “robotic quarterback” uses Windows, Azure ecosystems to solve developer dilemmas—and cut a year off time to market

West Virginia University added another trainer to its football team last year. The team’s wide receivers have benefited a lot from the additional help, and so have the defensive backs, kick returners, and “everyone else who has the potential to touch a football,” according to Coleman Barnes, the Senior Associate Athletics Director and Chief of Staff for West Virginia University (WVU) Football. 

Indeed they have. In the 2021 season, WVU’s dropped passes fell by 53 percent to just 4 percent, compared to the 2020 season, and Barnes says the new trainer deserves a share of the credit.

That new trainer is the Seeker, which designer and manufacturer Monarc calls “the world’s first robotic quarterback.” Monarc is a Dallas-based sports robotics and software company. And the Seeker is far more than a traditional football-throwing machine. It combines AI, the Internet of Things (IoT), programmability, and more to do what Monarc says no traditional football-throwing machine has ever done: enable players to train alone.

At WVU, defensive backs use the Seeker for rotational drills, catching balls overhead; wide receivers practice catching over-the-shoulder throws. WVU players and coaches easily program the Seeker for punts, kickoffs—any type of play they want. They select where the ball should go relative to the player (thanks to a wearable position tracker), how fast, at what arc, and with how much hang time. 

“We can incorporate our whole playbook into the Seeker,” says Barnes. “Anyone can work on whatever they want, whenever they want, and because the Seeker operates automatically, players can practice before or after practices and games without needing someone to manage the device.”

If team managers and players are passionate about the Seeker, perhaps it’s because the robotic quarterback fundamentally changes training in football. “In other sports, players can train alone, but not in football,” says Igor Karlicic, who cofounded Monarc along with Bhargav Maganti. 

“The need is so obvious; most people are shocked that no one had done this before,” adds Maganti.

To meet that need, mechanical engineers Karlicic and Maganti had to answer a lot of questions, including what technologies they’d use to build what became the Seeker. At first they experimented with Ubuntu and Raspberry Pi OS (Raspbian), but ran into several hurdles very quickly.

“Our software team is pretty small,” says Maganti, referring to Connor Early, the software engineer who joined Monarc to help bring the Seeker to life. “And there are so many development aspects to this project. We needed a development platform and operating system that would automate all the tasks we wouldn’t have time to build ourselves. Plus we wanted a mature, reliable platform with a broad set of tools with which we could be productive immediately. On the other hand, we wanted the greatest flexibility to mix and match technologies and adopt new ones as they became available.”

They wanted Windows for IoT. 

“With Windows for IoT, we came to market up to a year sooner than we could have with Raspbian,” says Karlicic. 

A lot of factors combined to make that possible. “We could get all the UI items, buttons, and forms we wanted, then prototype and develop as quickly as possible,” says Early. “Windows also meant Visual Studio, with which we were already familiar, further speeding and enhancing our software. We didn’t always have Seeker hardware to test on—but we didn’t need it or need to wait for it. With Windows, we could test it on a PC.”

Another big benefit of using Windows for IoT, according to the two founders, was device compatibility. “We had no issues with driver support, which made our hardware choices easy,” says Maganti. “As we added modules—for Wi-Fi and cellular, for example—the drivers were readily available. And as we expand the Seeker, device compatibility in the Windows ecosystem expands too; it’s just not a problem.”

In response to their continued product development and expansion, the company adopted Azure for IoT and other Azure services. “Azure wasn’t on our minds at first,” says Early. “But it’s become a crucial part of our product. For example, we can update everything from our software to individual calibration values remotely. And Azure IoT has a range of features that we’re now thinking of adopting, ensuring a smooth roadmap for the Seeker’s future.”

By choosing Windows and Azure, Monarc also had immediate interoperability with Microsoft SQL Server and the Mobile Apps of Azure App Service, both of which feature in the Seeker’s tech stack. Through this arrangement, the team can operate the Seeker in connected and nonconnected mode, storing practice data locally when the Seeker is not connected to the internet and uploading the data to the cloud when it is. That interoperability also extends to Linux, which runs in containers on the Windows ecosystem. When Monarc is ready, it can adopt EFLOW for Linux interfaces and workloads.

“So with Windows, we’re not missing out on anything,” says Karlicic.

Early points to the great breadth of documentation available for C# and the Windows ecosystem. “There’s a large history and a lot of resources out there on how people have written apps for previous versions of Windows,” he says. “Whenever I have a question, it’s easy to get an answer and keep on going. That’s been a huge timesaver for us.”

It’ll continue to be a timesaver for Monarc as the company adds further Windows and Azure capabilities to the Seeker. And football players won’t be the only ones to benefit.

“Beyond evolving our current platform, there are other markets that could benefit from our technology,” says Maganti. “The Windows and Azure ecosystems are broad and flexible enough to support us as we grow.”