AR/VR
XR Interview Prep: Questions, Patterns, Pitfalls
An XR engineer at Meta Reality Labs earns USD 200-300 thousand per year. Valve hires fewer than 10 XR engineers globally per year. Niantic seeks specialists in SLAM and computer vision. The market is small, expectations are high - but the required knowledge is narrow. Going deeper than most candidates is a realistic goal.
- Meta Reality Labs: 1000+ XR engineers, focus on Quest and Horizon Worlds
- Apple Vision Pro team: SwiftUI + RealityKit + visionOS - iOS plus XR depth required
- Niantic (Pokemon GO AR): georeferenced AR, CV, SLAM on mobile devices
- Unity XR: XR development tooling - both engine expertise and XR depth needed
Rendering Questions: Latency, Stereo, Artifacts
An interviewer at Meta XR or Valve asks questions not to test memory - but to probe systems thinking. "Why does motion sickness occur at 20ms latency but not at 10ms?" This question touches vestibular-visual conflict, physiology, and system requirements all at once.
**Foveated rendering in interviews** - implementation details are frequently asked. Two types: fixed foveated (fixed quality zone, no eye tracking) and eye-tracked foveated (zone follows gaze). Key nuance: eye tracking adds ~2-5ms of processing latency, and this delay must be compensated by predicting gaze direction. Otherwise the user notices a 'lagging' quality zone.
Why do VR displays prefer OLED/microLED over LCD?
Tracking Questions: SLAM, IMU Fusion, Drift
**Tracking is a favorite topic** for interviewers at Meta Reality Labs, Niantic, Qualcomm XR. The reason: it sits at the intersection of computer vision, sensor fusion, and systems programming. A weak answer: "cameras track the head". A strong answer: VIO plus SLAM plus loop closure, explained demonstrably.
**Loop closure** is a critical concept for VIO interviews. When the headset returns to a known location, the algorithm recognizes it (place recognition) and corrects accumulated tracking error. Without loop closure: walk a circle around the room and the starting point drifts 5-10cm. With loop closure: return to the same point within 1-2mm accuracy.
Why does XR tracking use a Kalman Filter combining IMU and camera data?
UX and Interaction Questions for XR
UX questions in XR interviews are often underestimated. But at companies like Meta, Apple, and Unity they are asked alongside technical questions. "How would one design an interface comfortable in MR?" - this tests understanding of spatial UI constraints.
Why is head-locked UI (following the head) undesirable in XR?
System Design for XR Applications
"Design the architecture of a multiplayer XR application" - a classic system design question for senior XR positions. Key differences from standard multiplayer: synchronizing 6DoF transforms instead of 2D positions, voice proximity (sound fades with distance), and body IK for an avatar from only head and hand data.
**Codec Avatars (Meta)** are neural network avatars that reproduce facial expressions and body motion from minimal sensor data. Trained on 3D scans of real people. Result: a realistic avatar with accurate facial expressions without a face-mounted camera. This direction is actively evolving - the next generation of XR social presence.
For an XR interview it is enough to know the Unity or Unreal API
Hardware constraints are expected knowledge: photon latency, tracking pipeline, stereo rendering, memory bandwidth on mobile GPU
XR is systems engineering. Misunderstanding latency or memory constraints leads to a product that causes motion sickness or runs at 30fps instead of 90fps. Interviewers test exactly this systems understanding.
What is dead reckoning in the context of XR multiplayer?
Key ideas
- Photon latency <20ms is a physiological requirement, not an arbitrary target
- VIO = IMU (fast) + Visual (accurate) + Kalman Filter (optimal fusion)
- Head-locked UI: only for critical HUD; world-locked or hand-locked is preferred
- XR multiplayer: 6DoF sync + spatial audio + body IK from three points
- System design questions test hardware constraints knowledge, not just API familiarity
Related topics
XR interviews span the full stack - from display physics to UX patterns.
- XR System Architecture — Foundation of all technical questions
- XR Performance — Profiling and optimization - top topic for senior positions
- Tracking and SLAM — Tracking algorithms - mandatory for deep technical questions
Вопросы для размышления
- How would one explain the difference between motion-to-photon latency and end-to-end latency in an XR system to an interviewer?
- Design a body IK system for a multiplayer XR avatar: what data is needed, what algorithms to use?
- How can an XR application be tested for motion sickness before user testing begins?