What Muhammad demonstrated

Built a solid modernization pass for a legacy fleet-maintenance service: fixed the wear calculation so nearly-worn cars are flagged, handled missing service readings safely, corrected the nightly report’s average and crash behavior, and repaired a hidden mileage-conversion bug in the helpers. Also documented the review process in NOTES.md and described a data-driven breakdown-risk analysis that prioritizes km-since-service and driving intensity over total mileage.

• Leakage-Free Feature Engineering
• Honest Model Evaluation
• Directed AI Prompting
• Verification and Audit Discipline