Nakul Armarkar completed Machine Learning for Automotive Safety

Built a complete perception validation dossier for a synthetic Aurora-7 detector: defined detection correctly with IoU >= 0.5, scored precision/recall and the VRU false-negative rate, and then broke recall down by lighting, visibility, and class rarity to expose the dangerous night and occlusion failures. The final recommendation was evidence-based and internally consistent, pairing a daylight-only conditional release with specific SOTIF mitigations for the highest-risk VRU misses.

• Leakage-Free Detection Definition
• Safety-Critical Metric Selection
• Per-Condition Failure Triage
• SOTIF Hazard Framing