Hi — I’m Nabeel Hussain
I’m a high school senior in the Chicago suburbs building baseball analytics tools using Statcast-style
features and sequence modeling. NextPitchAI (PitchGPT v3) predicts the next pitch bucket from game context.
What I built:
• A multi-input neural network that combines recent sequence encoding + game context features
• A deployed web demo (Flask + Render + custom domain)
• Supporting ML artifacts (scaler / encoders) to reproduce inference reliably
What’s next (v4 ideas):
• Pitcher-specific arsenal priors (MLBAM ID) to sharpen predictions
• Better batter-side handling (including switch hitters) with training-time fixes
• Expanded outputs (pitch-type level, not just 5 buckets)