We built this because AOG events are still happening over one missing forecast
Founded in El Segundo, CA in 2025 by a team with backgrounds in aviation supply-chain operations and machine-learning forecasting.
The data to prevent most AOG events already exists
Aviation maintenance organizations generate enormous volumes of removal history, task card data, and flight cycle records every day. The patterns that predict a parts shortage are buried in that data.
The problem is that no standard MRO system was built to read that data as a forecaster would. Parts planners compensate with experience, spreadsheets, and safety-stock rules that overfit to the last crisis. The result is inventory that is wrong in both directions: overstocked on parts that do not fail and understocked on the ones that do.
Aerotrax was built specifically to close that gap. One job: read your fleet's removal history and tell your planners what to order before the shortage becomes a ground event.
People behind the forecast
David Bettenhausen
CEO and Co-Founder
Spent years running demand planning and supply-chain forecasting for maintenance operations before co-founding Aerotrax. Focuses on customer workflow and product direction.
Sarah Chen
CTO and Co-Founder
Built time-series forecasting systems for industrial equipment wear prediction, then applied the same methods to aviation component degradation data. Leads the forecasting engine and data infrastructure.
Marcus Webb
Head of Product
Comes from hands-on MRO line maintenance work, the perspective most forecasting tools have never actually consulted. Owns the product experience so planners see what matters, not just what is technically possible to surface.
Principles that guide product decisions
Accuracy over optimism
We show confidence intervals, not just point forecasts. A planner who trusts a range they understand is more useful to their operation than one chasing a number they cannot verify.
Planner-first design
We did not build a dashboard for analysts. We built a weekly report for someone who needs to make a purchase order decision by Thursday. The output is a ranked list, not another interface to learn.
Honest about limits
If your removal history is too thin to produce a reliable forecast for a specific part, we say so instead of inventing a number. Your data stays yours and is never used to train models for other operators.
Ready to see what your data reveals?
30-day pilot on your own removal history. No integration required to start.
Request a pilot