Train With an AI Tennis Coach: SwingVision

Train and improve your tennis game with the help of artificial intelligence.

About SwingVision

Company: Mangolytics
Founders: Swupnil Sahai and Richard Hsu
Mission: To provide pro-level metrics to all tennis players.
App launched: 2016 (on Apple Watch); 2019 (on iPhone and iPad)
Team size: 7
Most-used emoji: 🎾

When Swupnil Sahai started developing SwingVision, he had never written a line of code before. However, as an avid tennis player, he wanted to have access to the same stats and insights as professional players. This motivated him to create an app that could provide users with real-time feedback and analysis on their tennis performance.

SwingVision is a revolutionary virtual tennis coach and highlight-reel creator. By simply pointing your iPhone or iPad's camera at the court while playing, the app can analyze key aspects of your game, including serve speed, shot consistency, posture, and footwork. It can also generate personalized video highlights and identify areas for improvement. With upcoming updates, it will even serve as a line judge during matches.

Interview with Swupnil Sahai

SwingVision: From Zero to Hero

Q: SwingVision is one of the most advanced AI-powered sports apps, but you started with no coding experience. How did you overcome this challenge?

A: Yes, it's true. I taught myself Swift so that I could create an Apple Watch app to track my shots and keep score. When I was pursuing my PhD at Columbia University, Swift was released. It seemed more accessible than Objective-C, so I took the opportunity to learn coding on my own. Swift turned out to be an incredible first language for beginners.

Q: How did you transition the app from Apple Watch to iPhone?

A: After graduating, I worked on autonomous driving and gained expertise in computer vision and AI algorithms for object tracking. This experience inspired me to apply similar principles to tennis analysis. With the advancements in iPhone technology, especially the powerful bionic chips, I realized that bringing SwingVision to iPhone was feasible. The iPhone's camera capabilities opened up new possibilities for improving the app's functionality.

Q: What has been your biggest technical challenge?

A: About a year and a half ago, we faced a significant challenge while developing the AI model that tracks shots using a single camera. The conversion process from training the machine learning model on a computer to running it on an iPhone with Core ML was not working correctly. We spent weeks searching through online forums for a solution. Eventually, we found a simple fix: changing just one line of code in the Python script before conversion. It's incredible how such a small tweak made all the difference.

Q: Tennis pros Andy Roddick and James Blake are investors in SwingVision. How did that happen?

A: I have an interesting story about this. I met James Blake during an exhibition match in Texas, where Andy Roddick, John McEnroe, and Jim Courier were also present. After pitching my app to them, Andy showed the most interest. However, what really sealed the deal was when Andy couldn't find his iPhone and James tried to call it without success. I suggested using Find My iPhone to locate it. This impressed both of them, as they were unaware of this feature. This incident led to their involvement and investment in SwingVision.

Q: Where do you envision SwingVision heading in the future?

A: Currently, SwingVision is most popular among club players and adults who play competitively once a week. However, our goal is to make tennis accessible to everyone, regardless of skill level. We're starting to see more high school players and casual players using the app. Our focus is on teaching people how to play tennis effectively and automating aspects like scoring. Tennis has a unique scoring system, and there's a need for tools that can simplify the process. We aim to complement the existing infrastructure and make tennis more accessible to all.