AirSynthix — AI Crew-Scheduling SaaS for Airlines

The problem

Crew scheduling is one of the hardest problems in aviation. Flight-time limits, mandatory rest, qualifications, fairness, and last-minute disruptions all pull against each other — and the rules change by regulator and by country. Spreadsheets and manual rostering collapse under that complexity, especially for regional and charter airlines that don’t have a large planning department.

What I built

AirSynthix is a constraint-optimization engine built on OR-Tools CP-SAT, with Benders decomposition so it scales to large fleets. It encodes multi-regulation compliance as hard constraints — FAA Part 117/135, EASA ORO.FTL, Transport Canada — so an illegal roster simply can’t be produced. When disruptions hit, it re-optimizes on the fly while preserving the existing plan (warm-start), and it splits monthly scheduling into overlapping weekly windows to stay fast.

It ships as multi-tenant SaaS: each airline’s data is isolated, access is role-based, and the whole product can be white-labeled.

The result

What used to take days of manual rostering — and still left rule violations to catch by hand — becomes an optimized, compliant schedule generated in minutes.