Verified Credential Download PDF
What Nandish demonstrated
Built a complete modernization of the KM-Waechter service: fixed the wear calculation so nearly-due cars are flagged correctly, handled missing service readings without crashing the report, corrected the fleet-distance conversion bug in the helper layer, and added the missing regression test. Also delivered a data-driven breakdown-risk analysis that identified the real predictors in the fleet history data and documented the verification process clearly in NOTES.md.
- Leakage-Free Feature Engineering
- Honest Model Evaluation
- Directed AI Prompting
- Verification Discipline
Earn your own
Prove your skills on a real project
Pick a real company brief, complete it step by step, and walk away with a verified certificate like this one, graded by AI, shareable anywhere.
Start your own project →