Self-directed project
AI-Assisted Code Modernization with IBM
Junior Software Engineer · Software Engineering · September 2026
Based on a public IBM case study · not affiliated with IBM
Built and verified a full legacy-code modernization for KM-Waechter: corrected the service-wear calculation, handled missing readings safely, fixed the nightly report’s average and crash path, and repaired a quiet mileage-conversion bug in the helpers. Also added a breakdown-risk analysis that follows the fleet-history data rather than the obvious but unsupported mileage/age assumption, and documented the verification process in NOTES.md.
Graded against
- Fix and extend the repo with the agent35%
- Diagnose the bug20%
- Direct and audit the AI agent25%
- Explain it like an engineer20%
Passed · pass mark 60/100
What stood out6
- Fixed the hidden km_to_miles conversion bug in fleet_utils, which is the strongest sign they swept beyond the obvious tests
- Kept the 15000 km interval and 80% threshold unchanged while correcting the service logic
- Wrote a data-driven analyze.py that explicitly rejects the misleading odometer/age assumption and focuses on km_since_service, avg_daily_km, and load_factor
- Leakage-Free Bug Diagnosis
- Directed AI Code Review
- Acceptance-Driven Verification
The deliverable
The work I submitted18 tasks
10 tasks · 11 lines · 8 written answers · 143 words
- Division Predict73 characters
- Diagnose Line130 characters
- Scope Prompt66 characters
- Audit Findings169 characters
- Template Repo Click152 characters
- Bob Trial Click202 characters
- Exec Pitch645 characters
- Bob Bonus151 characters
- Repo URL58 characters
- Jupyterlite Code2 lines
- Division Predict1 word
- Diagnose Line10 words
- Scope Prompt1 word
- Audit Findings1 word
- Template Repo Click1 word
- Bob Trial Click1 word
- Exec Pitch99 words
- Bob Bonus29 words
Summarised on purpose — the submitted code and writing stay private so this page cannot be reused as an answer key. The full submission sits behind the verified certificate.