Yash Jangir completed Machine Learning for Automotive Safety

Built a perception-validation dossier that correctly locked the IoU-based detection rule, computed aggregate precision/recall and the VRU false-negative rate from a scored table, and then exposed the dangerous night and occluded slices where VRU recall collapsed. The analysis tied those failure modes to the real Uber ATG Tempe crash and showed the candidate can connect metrics, safety interpretation, and real-world incident review in one coherent workflow.

• Leakage-Free Detection Definition
• Safety-Critical Metric Selection
• Per-Condition Failure Triage
• Evidence-Based Real-World Mapping