Whenever you're ready, walk me through how you'd approach this.
Describe a time you analyzed baseball pitch data to predict whether a pitch was a ball or strike from pitch and FX features.
Explain how you defined the target, prepared pitch-level and FX features, selected and trained a classification model, and evaluated its performance. Discuss validation design, potential leakage across games or pitchers, class balance, error analysis, and how you would communicate the model's limitations. Include representative Python code or pseudocode for the modeling workflow.