Darshan Patil completed Machine Learning for Automotive Safety

Built a perception validation dossier that defined detection correctly at IoU ≥ 0.5, scored synthetic detections with precision/recall and VRU false-negative rate, and then sliced performance by lighting, visibility, and class rarity to surface the highest-risk misses. The final recommendation was internally consistent and evidence-based, tying a daylight-only operating restriction to the measured night-time VRU recall collapse and framing the risks as SOTIF functional insufficiencies with concrete mitigations.

• Leakage-Free Detection Definition
• Honest Metric Evaluation
• Per-Condition Safety Triage
• SOTIF Functional-Insufficiency Reasoning