Interview Questions › Product Design

Design an AI-human co-pilot interface for Air Traffic Controllers (ATC)

Design a safety-critical AI decision support system and radar display interface for Air Traffic Controllers managing 40+ simultaneous aircraft landings, takeoffs, and storm avoidance vectors during peak airport congestion.

Context

ATC controllers work under extreme cognitive stress. System errors risk catastrophic loss of life, while poorly designed AI creates "automation complacency" or alerts that controllers silence due to false alarms.

Clarifying questions to ask first

• Does the AI execute autonomous flight commands or provide advisory recommendations to the human controller? (Advisory with human authorization).
• What is the required system uptime? (99.9999% six-nines mission critical).

How to approach it

1. Analyze High-Stakes Cognitive Workload: Spatial 3D airspace visualization, high radio frequency voice congestion, rapid conflict resolution under 15-second decision windows
2. Design Human-in-the-Loop AI Collaboration Architecture:
3. 1. "Predictive 4D Trajectory Conflict Predictor" (Identifies collision loss-of-separation risks 8 minutes in advance with transparent reasoning)
4. 2. "Recommended Vector Solutions with 1-Click Approval" (AI proposes optimal heading/altitude alteration that resolves conflict without cascading into adjacent sectors)
5. 3. "Automated Pilot Voice Readback Verifier" (Listens to radio frequency and instantly flags if a pilot misunderstood an altitude clearance)
6. 4. "Dynamic Cognitive Workload Balancer" (Auto-detects controller fatigue/pupil dilation and recommends sector handoffs)

Framework: ATC Cognitive Workflow & Disaster Risks -> Conflict Prediction & Transparent AI Reasoning -> Controller UI & 1-Click Voice Clearance -> Fail-Safe Redundancies -> Safety SLAs

Sample answer

• AI must never be a black box in aviation; every vector recommendation must visually show the projected separation margin and storm clearance path
• Automated audio listening to pilot readbacks eliminates the #1 cause of runway incursions (pilot mishearing runway numbers)

1. MVP: 3D trajectory conflict alert radar, 1-click proposed resolution vectoring, real-time pilot radio readback speech-verification banner, and emergency wake-vortex collision avoidance alerts

Executive design: A mission-critical ATC co-pilot combining predictive 4D trajectory conflict detection, explainable vector resolution recommendations, and real-time radio readback speech verification.

What interviewers look for

• Safety-Critical Human-in-the-Loop Architecture
• Explainable AI Decision Recommendations (XAI)
• Radio Frequency Voice Readback Verification
• False Alarm Suppression & Cognitive Ergonomics