← All work

Flagship case study · Defense · Ground Control

Inventing a control language for machines that fly.

A ground control station for autonomous drone swarms in defense operations, where operators, analysts, and commanders share one screen and every click can affect lives.

01 · Context & business problem

Mission logic lived in code. Operators lived in fear of it.

We weren't redesigning. We were inventing a control language for machines that fly at 144 km/h, make real-time decisions, and operate in groups.
  • Prechecks were scripts. Mission-critical logic lived in code, with no interface for non-developers.
  • Missed steps meant aborts. A skipped logic step meant flight failures and wasted missions.
  • No swarm planning existed. Multi-drone synchronous missions could not be designed at all.

02 · My role & constraints

UX from zero, under field conditions.

I owned the UX end to end: research with tactical drone pilots, mission analysts, and defense engineers; the information architecture; and the three core flows. One operator set the bar: “Too many buttons. I'd hesitate instead of launch.”

  • Glove-friendly, daylight-usable, meaning field conditions, not desk conditions.
  • Works without training, because hesitation is a failure state.
  • Calm under stress, with sparse layouts, dedicated alert zones, reversible actions.

03 · Architecture

A layered command architecture before any screen.

Operator insights became design law: action always visible, destructive commands get timed confirmations, every flow interruptible and reversible.

The UX Command Stack: five layers, one home for every element

  1. Action Layer

    Instant commands and triggers.

    Launch / Abort / Payload buttons
  2. Mission Manager

    Plan, assign, and configure drone behavior.

    Mission flow, waypoint setup
  3. Swarm Status Layer

    Always-on awareness of each drone’s health and role.

    Drone cards, telemetry bars
  4. Visual Map Layer

    Spatial awareness, drone trails, and danger zones.

    Live map with overlays + icons
  5. Feedback & Alerts

    Communicates critical events, states, and errors.

    Error modals, state chips, confirmations

04 · Validation

Validated with the people who fly them.

Round 1: 6 drone engineers, 2 QA analysts. Round 2: live operations with 4 new operators and 3 mission commanders.

21 heuristic issues found & fixed 8 of 12 walkthrough ideas shipped pre-release 28 UI bugs + 14 UX flaws closed 0 launch-blocker bugs

05 · Results

The outcomes, as measured.

Task success rate in testing

Before 58%
After 92%

Time to build a mission flow

Before 19 min
After 8 min

Operator time to proficiency

Before baseline
After −50%
↓60% error clicks 98% QA confidence: “ready for live ops”

06 · What I would do differently

Simulate stress earlier.

The biggest lesson, designing for stress rather than just use, arrived through testing. Building even a crude swarm simulator first would have surfaced the hesitation moments before any operator felt them. The best interfaces are the ones you forget about mid-flight.