Passion project

Building & Improving Anduril's EagleEye UX in 24 Hours.

VR/AR/MR

Unity engine

Vibe Coding

White virtual reality headset above a soldier wearing a tactical helmet and sunglasses, with digital targeting graphics including radar with red and green markers and compass data overlaid on a nighttime outdoor scene.

Summary

In 24 hours, using Claude AI as an assistant, I designed and built a working mixed reality (MR) prototype of Anduril's “EagleEye” goggles in Unity with zero prior game, VR, or coding experience. Users place friendly and hostile markers around a room, see markers through walls, and track them on a 360° compass and mini-map.

Impact

By my own comparison, my UI improves on Anduril’s demo HUD to support faster decisions under stress, including:

  • User orientation: A four-direction compass, large heading in degrees, and proximity ring show which way users are facing and how far away things are, at a glance.
  • Threat information: UI automatically shows the number of threats, easily identifiable hostile markers, and the freshness of marker intel.

NOTES

  • This project is a rough experiment with vibe coding, VR/AR/MR, and game engines for my own personal design growth. There's no implied value for actual defense applications or affiliation with Anduril.
  • The work took three 8-hour days, ~24 hours in total.

ROLE

Owned design of this project, including:

Research + UI design + UX design + Interactive prototyping

COLLABORATORS

Claude (AI): Code + Guidance

INDUSTRY

Technology + Defense

TIMEFRAME

Q3 2026 (24 hours)

Context

Anduril's EagleEye demo depicts mixed reality (where digital things appear over and interact with the real world) to show friendly and hostile positions in real time. I wanted to see if a hobbyist could recreate its core features on a regular headset.

Problems

Anduril's UX

Anduril’s public demo has four usability gaps:

-

▪

Weak orientation.

The compass only shows North, the heading is easy to miss, and nothing helps calculate distance.

-

▪

Manual counting.

Users must count threats by hand, costing time and focus.

-

▪

Color-only markers.

Friendlies and hostiles are identical squares that differ only by color, so they are easier to confuse under stress.

-

▪

Missing intel.

Markers don’t show distance or how old the intel is.

My constraints

My goal was to improve my vibe-coding skills and ship a user-centric VR/AR/MR experience. Three constraints made that hard:

-

▪

No VR or game experience.

I had never used Unity (game engine used to build VR apps), designed for VR/AR/MR, or written code before.

-

▪

Tight timeline.

The build had to fit in about 24 hours, across three 8-hour days to simulate a quick startup environment.

-

▪

Hardware limits.

Real-time see-through-walls is beyond the Quest 3, and true object detection would take too long to build.

Approach

Anduril's UX

To improve on Anduril's HUD, I designed mine around two principles:

-

▪

Do the work for users.

Users under stress shouldn’t have to calculate or interpret anything, so make everything obvious and automatic.

-

▪

Orient at a glance.

Give users clear direction, heading, and range cues so they stay on target and give better callouts.

My constraints

Three ideas kept the build on track:

-

▪

Claude codes, I design.

Claude handles the technical work, and I own the design.

-

▪

Small scope.

Drop features that are too advanced to keep the build under 24 hours.

-

▪

Test early and often.

Adjust the design after each test build to avoid wasted work.

Text conversation showing a user saying they have no coding, game engine, or VR experience, followed by a response outlining next development steps: leading from here, focusing on the minimap and settings overlay, and checking code related to the Unity project and building blocks for calibration.

Process

Hours 0–0.5

Claude Setup

I prompted Claude to act as my lead designer and engineer.

“Help me design, code, and test a simple but impressive VR/AR goggle app on my Meta Quest 3. I have no coding experience or game design experience so you will take on the role of lead game designer and software engineer and instruct me what to do in detail. The project will involve some sort of UI from Figma. I need your help coming up with an app idea on the Quest 3 goggles that would interest hiring managers for a job description at Anduril. I have thought about trying to make some sort of object tracking system to mimic their Eagle Eye product. I prefer to build the project in Unreal Engine but Unity is okay too; if there is another tool that's better please advise me. I would like the project to be done in 3 days or less if possible.”

Hours 0.5–2

Gathering Requirements

Real-time see-through-walls and true object detection were out of reach, so I cut the scope to:

First-person view aiming a gun with a laser sight at a windmill in a grassy, shrub-filled landscape during dusk, with a circular digital compass and map display at the bottom left showing direction and terrain.

Anduril's public EagleEye demo: Marking locations

Thermal vision view of an outdoor area with two shipping containers and a vehicle. Three human figures are outlined, two in cyan holding weapons near the center container, and one in red near the left container door. A radar display shows directional markers and targets. A translucent overlay on the right displays the label 'Ghost-X' with heading and altitude details and a small thermal map of the containers and figures.

Anduril's public EagleEye demo: See-through objects (not within the scope of this project or the capabilities of the Quest 3)

I sized the HUD at 1700 x 1500, inside the headset's 2064 x 2208 view, to keep elements near the center of vision.

Hours 2–4

Designing in Figma

Figma Make's three prompted attempts came out too light, too small, too detailed, and poorly grouped.

Modern living room with two gray sofas facing each other, a round black coffee table with flowers on a light rug, two windows with white curtains, a large TV on dark shelving, and a hanging ceiling lamp.

The first Figma Make iteration. The HUD is too light and small.

Sentinel system idle screen showing compass heading 023 degrees north, power at 72%, zero anchors, local time 18:39 on Friday, July 31, 2026, and a local radar map with no detected objects.

The second Figma Make iteration. Better but legibility is poor.

Dark futuristic heads-up display with compass showing north and northeast, central crosshair, time 18:39, date Fri, Jul 31, battery icon at 72%, and a radar with a dot in bottom right.

The final Figma Make iteration. The HUD is still too light and small, and it's not grouping elements together logically.

AI designs only help if they are designed well, so I designed the HUD by hand and connected Figma to Claude.

Dark interface showing a radar-like compass with north at 0 degrees, a red triangular marker labeled 'HOSTILE 10m' and a green circular marker on the radar and top bar, and a small red trash bin icon below the red marker.

My initial hand-designed Figma HUD. It changed after testing the first prototype.

Hours 4–6

Unity Setup

I set up Unity and Meta's VR toolkit with Claude, which gave us passthrough (seeing the room through the headset) and hand tracking.

User interface of a Building Blocks library for XR development showing search results for 'passthrough' with thumbnails of XR features like Passthrough Camera Access, Passthrough Camera Visualizer, Real Hands, Occlusion, and Passthrough Window, each illustrated with characters wearing VR headsets or interacting with virtual objects in a stylized room.

Hours 6–14

Building Core Functionality

Claude suggested building the compass first, then the tagging functionality: place markers on real walls, keep them fixed, and dim them when blocked. By day two, I had a simple prototype to test.

Unity 3D editor view showing a simple scene with a flat ground and gradient sky, centered on a directional compass indicator marked NE 000°.

Early version of the compass in Unity

Testing showed it was too tedious: adding a marker meant aiming my head, aiming my finger, and pinching. Removing markers was often painful too.

Dark interface section labeled Center Cursor with a dashed purple border containing five circular outlines: a grey dashed circle, three white dashed circles with a plus icon below the second, a red triangle inside a black square overlapping the third circle, a green circle inside a black square overlapping the fourth, and a red trash can icon below the fifth circle.

Default, Add, Select marker, and Remove cursor states. Too complicated in hindsight.

Hours 14–20

Rework

I replaced that aiming method with a menu that stays on screen: aim at a marker, then point at Add or Delete, and pinch. I cut the cursors from four to two, moved the tally above the compass, and canceled plans to add the top compass because it was redundant.

Digital compass display showing direction as North 0 degrees with concentric circular dial marked with N, E, S, W and a shaded arc indicating heading; includes two small buttons with red triangle and green circle symbols, and a gear icon button on the right.

Redesigned HUD layout

A dark interface panel labeled 'Center Cursor' with a purple dashed border. Inside, two dashed circles are displayed: a gray one on the left and a white one on the right. Below the white circle is a black rectangular label showing '10m 22s ago' in digital-style text.

New, simplified center cursors

Hours 20–24

Test and Polish

Testing led to smaller fixes:

Final Build

The app lets you place hostile and friendly markers in your room, see them through walls, and keep a count of each. Hovering shows a marker's distance and age.

Adding and removing a Hostile Marker on my cat. She deserves it.

Markers change opacity when losing line of sight to prioritize “urgent” threats

A compass with a large heading in degrees and a 360° mini-map show where markers are. Markers pair color with shape: sharp and red for hostile, round and green for friendly. A central dotted line helps users align their view with distant objects. The middle proximity ring helps users estimate distance.

Digital compass interface showing 0 degrees north with a red triangle marker aligned at north and a green circle marker to the east, on a dark background.

See all marker information in one place at a glance

Markers rotate around the user and change distance as the user moves

A demo menu adds dozens of random markers for testing, and one pinch removes them all.

Adding dozens of random markers for quick demonstration purposes

Removing all world space markers with one pinch

Impact

By my own comparison, my UI improves on Anduril’s demo HUD to support faster decisions, including:

+

Superior compass orientation.

Anduril’s compass labels only North. Mine labels all four directions, so users can orient at a glance.

+

Instantly readable heading.

Mine shows a large heading in degrees, so users can call out an exact bearing. Anduril’s is easy to miss.

+

Easier alignment and range.

A proximity ring estimates distance, and a dotted line helps users align their view with distant objects.

+

Automatic threat count.

Mine counts friendly and hostile markers automatically. Anduril’s has no tally, so users must count by hand and lose focus.

+

Intuitive shape and color.

Anduril’s markers are squares that differ only by color. Mine also differ by shape, likely making them easier to tell apart under stress.

+

Intel age and distance.

Hovering over a marker shows its distance and age, so users can judge the threat and how fresh the intel is instead of guessing.

Left: Anduril's public demo
Right: My design

Next Steps...

Here's what I'd do differently in hindsight:

Contact

Thank you for viewing my work. Schedule a call to go deeper on my workflow and design decisions.