About the role
The next way people live with their agent is hands-free: on glasses, on the watch, in a headset. It sees what you see by consent, speaks quietly, and acts, with you in control. This is early, genuinely hard, and mostly unbuilt, which is the appeal. You will be one of a small number of people anywhere with real shipped experience in this, working out what the interaction model even is.
What you’ll do
- Build 🤫 Agent One for visionOS (ARKit, RealityKit, Reality Composer Pro) and for Meta (Quest, Meta Spatial SDK, and the glasses Wearables Device Access Toolkit)
- Design glanceable, voice-first experiences for watchOS and Wear OS - the agent on your wrist
- Make spatial UX feel calm and natural: gaze, gesture, hand-tracking, and audio-first interaction
- Keep it consent-first on the most sensitive sensors there are: on-device perception, minimal disclosure, a receipt for anything the glasses see or share
- Push frame rate, latency, and thermals so long sessions stay comfortable and effortless
What we need to see
- You have shipped a spatial or wearable experience: visionOS, Quest, ARKit or ARCore, watchOS or Wear OS
- Fluency with the constraints that make this hard: latency budgets, thermal limits, and comfort over a long session
- You can design an interaction and not just implement one, because there is no established pattern to copy here
- Judgement about what should be hands-free at all, rather than porting a phone app onto a face
Nice to have
- RealityKit, Reality Composer Pro, or the Meta Spatial SDK specifically
- Audio-first interaction design, since much of this is voice rather than visual
- Computer vision or sensor fusion
What winning looks like
- Spatial and wearable experiences shipped to real users
- Frame rate, latency, and comfort within target on device
- Adoption and session quality on glasses, watch, and headset
Where and how we work
In the office together five days a week, in any of these cities. Remote-friendly around your family, arranged one person at a time.