Mohammad Awais Nijamoddin Pinjari completed Machine Learning for Automotive Safety
What Mohammad demonstrated
Built a perception validation dossier that defined detection by IoU threshold, scored a synthetic detections table, and separated aggregate performance from the safety-critical VRU false-negative rate. The analysis then broke recall down by lighting, visibility, and rarity, correctly exposing the night and occluded slices as the highest-risk failure modes.
- Leakage-Free Detection Definition
- Safety-Critical Metric Selection
- Per-Condition Failure Analysis
The challenge Mohammad passed
Machine Learning for Automotive Safety
In this German industry project, you build a five-section dossier: score a frozen detector on a synthetic detections table (proportions grounded in the public KITTI/nuScenes benchmarks) in Pyodide, find where pedestrian recall collapses, write the ISO 21448 SOTIF case, then sign the verdict against the real Uber Tempe 2018 crash.
- 2-3 hours
- Advanced
- Your role: Perception Validation Engineer (ADAS Safety Sign-off)
- Rated 4.2 by 62 students
- 300+ students started it
- Autonomous Driving & ADAS
- Functional Safety
- Testing & Validation
- Computer Vision
What you will do
- 01The sign-off is on your desk
- 02What does the car actually see?
- 03Score the detector
- 04Hunt the misses, build the case, make the call
- 05Read it against a real crash, then submit
Verified by ProoV · c88a378a8ddc306d…
