Meher Choudhary completed Machine Learning for Automotive Safety

Built a perception validation dossier end to end for a synthetic Aurora-7 detector: defined detection at IoU >= 0.5, scored precision/recall and VRU false-negative rate, sliced recall by lighting/visibility/class to surface the worst safety gaps, and translated those gaps into ISO 21448 SOTIF chains with concrete mitigations. The final release recommendation was evidence-based and tied directly to the measured night and occlusion failures, including a real-case comparison to the Uber Tempe crash.

• Leakage-Free Detection Definition
• Safety-Critical Metric Selection
• Per-Condition Failure Triage
• SOTIF Functional-Insufficiency Reasoning