AI-Assisted Flight Scheduling
In collaboration, built a predictive model that helped identify training slots to keep Air Force pilots compliant with FAA and USAF flight hours and regulation.
Product Designer, UX strategist · 2024–2025 · Sanitized
Project overview
- My role
- I owned research and design for building out NOVA's MVP, which unified permissions, privileges and required paperwork.
- My team
- I worked with USAF pilots, product teams, MIT AI Accelerator partners and 8+ partner government agencies at Kessel Run.
- What shipped
- The work covered NOVA's MVP and a predictive model that identifies open training slots. This public account describes the workflow; release scope is not disclosed.
- What changed
- Permissions, privileges and required paperwork came into one place, and schedulers could see training and airspace conflicts on one timeline before a schedule went out.
- What I can share
- Sanitized case study. No internal interfaces, operational data, or restricted workflows are shown. The schedule is recreated with notional data.
Context
NOVA was Kessel Run's replacement for Patriot Excalibur (PEX), the legacy flight-scheduling system pilots and schedulers relied on. Permissions, privileges and required paperwork were tracked separately from the schedule, so NOVA's MVP set out to bring them into one place.
Air Force pilots have to fly and train enough to stay compliant with FAA and USAF flight-hour rules. Schedules were built across Excel, Word, legacy software and separate databases, and every scheduler worked with them differently.
When a sortie collided with a training block or closed airspace, someone had to deconflict the slot. In collaboration with MIT Phantom Fellows, we built a predictive model that identifies open training slots, and designed how schedulers see and resolve conflicts on one timeline.
As Product Designer and UX strategist at Kessel Run, I owned research and design for NOVA, the Air Force's replacement for its legacy flight-scheduling system. Over two years, I worked with pilots, product teams, MIT AI Accelerator researchers and 8+ partner government agencies, from discovery through the MVP:
- Built out NOVA's MVP, which unified permissions, privileges and required paperwork in one place
- Interviewed 20+ USAF pilots to understand how they scheduled, trained and stayed compliant across Excel, Word, legacy software and databases
- Maintained 20+ regulatory manuals so scheduling reflected current FAA and USAF flight-hour rules
- Partnered with the MIT AI Accelerator on a predictive model that identifies open training slots and keeps pilots compliant
- Designed the deconfliction workflow shown above, where a scheduler resolves training and airspace conflicts before a schedule goes out
