SSIM Schedule Validation App

Challenge

Schedule files flowed to the maintenance-planning system (CAE AMDM) and the crew system (Merlot) with no validation gate. Bad airport codes, unknown aircraft types, duplicate flights, and local-versus-UTC mistakes surfaced downstream, in vendor systems, after the fact. The pipeline also had to keep flowing through the flight-ops platform migration with strict IATA SSIM conformance.

Approach

• ▸Compared an AWS Lambda proof of concept against a MuleSoft proof of concept and chose the architecture; ran UI mockups through a feedback loop with the schedule planner who uses it daily.
• ▸Specified the validation rules: airport and aircraft-type cross-checks against the vendor master lists, duplicate-flight prevention per flight number and UTC day, effectivity-date validation, and UTC date-cutoff trimming with a review-and-approval step.
• ▸Added local-time auto-detection with UTC conversion for AMDM delivery, plus a CSV export carrying local date, time, and service type.
• ▸Opened a second delivery path: parsed CSV to the crew system with REMOVED-flag handling, and an AMDM parity check that compares delivered legs to processing results and emails on mismatch.
• ▸Moved the infrastructure from SST to Pulumi; added QA automation for the validation UI; drafted the stateful rework that diffs new uploads against the delivered schedule and sends updates instead of full reloads.
• ▸Applied working knowledge of IATA SSIM Chapter 7 (fixed-width 200-character records, Record Type 3 period semantics, weekday encoding, UTC offsets) and built Python tooling for generation, validation, and analysis.

Outcome

• Every schedule file is validated before it reaches either vendor system; delivery is auditable end to end.
• Documented a structural constraint (period-based records cannot express a time-of-day cutover) that directly shaped the migration cutover-window plan.
• Unblocked larger schedule files early with an upload-limit release in April 2025 while the full product was built.