1
We began by brainstorming the core features and functionality of BallsEye, focusing on precision and usability.
2
Next, we collected training data and applied machine learning techniques to detect line calls accurately.
3
We integrated the model into a lightweight app with a clean UI that could run smoothly in real time.
4
Finally, we tested and refined the app, iterating on performance until it was reliable for real matches.