Tvisha Guduru completed Machine Learning for Automotive Safety

Built a perception validation dossier end to end for a synthetic Aurora-7 detector: defined detection with an IoU >= 0.5 match rule, computed aggregate and VRU-specific metrics, exposed the night/occluded recall collapse, and translated those findings into a conditional release recommendation. The work shows strong safety-oriented evaluation habits and a clear understanding that VRU recall, not headline accuracy, governs the release decision.

• Leakage-Free Detection Definition
• Honest Safety Metric Evaluation
• Per-Condition Failure Triage
• Evidence-Based Release Decision