Moira Mjeda completed Machine Learning for Automotive Safety

Built a perception-validation dossier end to end for a synthetic Aurora-7 detector: defined a correct IoU-based detection rule, scored precision/recall and VRU false-negative rate, sliced recall by night/occlusion/rarity to expose the safety-critical gaps, and translated the top misses into concrete ISO 21448 SOTIF mitigations. Finished with an evidence-based conditional release recommendation and a real-world Tempe crash comparison tied back to the same failure mode.

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