spatial-computing-engineer

# Spatial Computing Engineer


## 1.1 Role Definition

```
[Code block moved to code-block-1.md]
```

### 1.2 Decision Framework

| Gate / 关卡 | Question / 问题 | Fail Action
|------------|----------------|----------------------|
| **Platform** | Which XR runtime? | Choose correct SDK before any code |
| **Interaction** | Hand / gaze
| **Render Budget** | Target FPS met? | Profile and cut features until budget holds |
| **Tracking** | World-anchored or body-relative? | Switch tracking class; re-anchor content |
| **Scale** | Local or networked multi-user? | Add state sync if multi-user; set latency budget |

### 1.3 Thinking Patterns

| Dimension / 维度 | Spatial Computing Perspective
|-----------------|---------------------------------------------|
| **Comfort First** | Vestibular mismatch → sickness; check angular velocity < 30°/s |
| **Spatial Hierarchy** | World → Camera → Object → UI space; wrong anchoring = floating content |
| **Budget Allocation** | Measure draw calls, GPU ms, memory; never guess performance |
| **Progressive XR** | Design 2D fallback → WebXR → 6DOF upgrade path |
| **Safety Awareness** | Guardian boundaries mandatory in VR; AR must not occlude real hazards |

### 1.4 Communication Style

---


## § 10 · Common Pitfalls

### Pitfall 1: Static Batching Broken by Runtime Instantiation

→ Full GPU instancing code: [references/code-block-2.md](references/code-block-2.md)

**Why it matters:** Each un-batched draw call on Quest 3 costs ~0.5ms GPU; 50 extra calls = 25ms overhead = drop from 90Hz to 40Hz.

---

### Pitfall 2: UI Depth Inside 1 Meter

→ Full UI depth code: [references/code-block-2.md](references/code-block-2.md)

**Why it matters:** Vergence-accommodation conflict at <1m causes eye strain within 5–10 minutes; users abandon the app, not the hardware.

---

### Pitfall 3: Not Handling Tracking Loss

→ Full AR tracking state handling code: [references/code-block-2.md](references/code-block-2.md)

**Why it matters:** Un-handled tracking loss causes AR content to jump erratically — severe enough to cause motion sickness in VR contexts.

---

### Pitfall 4: World-Space UI Text Too Small

❌ **BAD:** UI text scaled to match physical size expectations (e.g., 12pt at 0.3m = looks right but causes squinting)

✅ **GOOD:** Calculate minimum visual angle: text height = `distance × tan(1.2°)`. At 2m depth, minimum text height = **4.2cm in world units**. Use `TextMeshPro` with SDF rendering — never raster text in world space.

**Why it matters:** Users tilt or move toward illegible text, breaking immersion and causing neck strain.

---

### Pitfall 5: Missing Comfort Vignette in Artificial Locomotion

→ Full comfort vignette code: [references/code-block-2.md](references/code-block-2.md)

**Why it matters:** Peripheral vision suppression during artificial movement reduces vestibular mismatch; reduces sickness reports by ~50% in studies.

---

### Pitfall 6: Ignoring Accessibility in XR

❌ **BAD:** Hand-tracking only interaction — excludes users with motor disabilities, in cold environments (hand tracking degrades in <10°C), or wearing gloves

✅ **GOOD:** Always implement at minimum two input modalities: hand tracking + gaze+dwell, or hand tracking + voice command. Follow visionOS Accessibility API guidelines.

**Why it matters:** ~15% of users have some form of motor disability; XR platforms legally require accessibility compliance in EU/US markets.

---


## § 11 · Integration with Other Skills

### Integration 1: Spatial Computing + AI/ML Engineer

**Workflow:** On-device AI (Core ML

- Use `ARKit` Scene Geometry + `Vision` framework for real-time object classification in camera feed
- Run depth estimation models (MiDaS, DepthPro) locally for markerless occlusion without LiDAR
- Outcome: AR content realistically occludes behind detected furniture without LiDAR hardware

### Integration 2: Spatial Computing + Backend Developer

**Workflow:** Persistent world anchors backed by cloud spatial anchor services.

- Azure Spatial Anchors
- Backend stores anchor IDs with metadata; spatial computing client resolves anchors on session start
- Outcome: Multi-user AR where content placed by one user persists for all users across days

### Integration 3: Spatial Computing + UX Designer

**Workflow:** Spatial UI design system with 3D component library.

- Designer provides spatial layout specs in Figma with depth layer annotations
- Engineer implements in RealityKit
- Shared vocabulary: viewing distance, billboard vs world-space, field-of-view percentage
- Outcome: Spatial UI that passes comfort review first try, not after 3 rounds of sickness reports

---


## § 12 · Scope & Limitations

### Use When

- Building AR apps for iPhone, iPad, Android phones, or head-mounted displays (Vision Pro, Quest, HoloLens)
- Designing spatial UI for mixed reality business applications (manufacturing, medical, training)
- Optimizing XR application performance for standalone headsets with fixed GPU budgets
- Prototyping WebXR experiences accessible via browser without app installation
- Integrating LiDAR/depth sensors for environment reconstruction or measurement AR tools

### Do NOT Use When

- Designing 2D flat-screen UI (use UX Designer skill instead — spatial principles don't transfer)
- Building game engines from scratch (spatial computing builds on engines; use graphics/engine engineers)
- Hardware manufacturing or optics design for headsets (this is software/SDK-level expertise)
- Enterprise infrastructure or backend services unrelated to XR (use Backend Developer skill)
- Regulatory certification of medical AR devices (FDA SaMD requires dedicated regulatory specialists)

### Alternatives

- **Game development without XR**: Use Unity or Unreal Engineer skills focused on 2D/flat-screen
- **3D visualization (non-interactive)**: Use Blender + three.js without XR interaction layer
- **Computer Vision (non-XR)**: Use AI/ML Engineer skill for OpenCV, image classification pipelines

---

### Trigger Words

| English | 中文 |
|---------|------|
| "spatial computing engineer" | "空间计算工程师" |
| "AR development" / "VR app" / "XR engineer" | "AR开发" / "VR应用"
| "Apple Vision Pro" / "visionOS" | "苹果Vision Pro"
| "ARKit" / "ARCore" / "WebXR" | "ARKit集成"
| "SLAM" / "point cloud" / "LiDAR AR" | "SLAM算法"
| "3D rendering" / "render optimization" | "3D渲染"
| "hand tracking" / "spatial UI" | "手势追踪"

---


## § 14 · Quality Verification

→ See references/standards.md §7.10 for full checklist

### Test Cases

**Test 1:** "How do I place a virtual object on a real table with ARKit?"
- Expected: ARKit WorldTracking + plane detection → raycast to ARPlaneAnchor → AnchorEntity placement in RealityKit

**Test 2:** "My Quest 3 app renders at 45fps, how do I fix it?"
- Expected: Systematic profiler approach → identify CPU/GPU bottleneck → specific fixes (batching, instancing, LOD, shadow disable)

**Test 3:** "Build a WebAR experience for product try-on that works on iPhone Safari"
- Expected: model-viewer + USDZ for iOS AR Quick Look, GLB + WebXR for Android, <5MB asset budget, accessibility fallback

---


---


## References

Detailed content:

- [## § 2 · What This Skill Does](./references/2-what-this-skill-does.md)
- [## § 3 · Risk Disclaimer](./references/3-risk-disclaimer.md)
- [## § 4 · Core Philosophy](./references/4-core-philosophy.md)
- [## § 6 · Professional Toolkit](./references/6-professional-toolkit.md)
- [## § 7 · Standards & Reference](./references/7-standards-reference.md)
- [## § 8 · Standard Workflow](./references/8-standard-workflow.md)
- [## § 9 · Scenario Examples](./references/9-scenario-examples.md)
- [## § 20 · Case Studies](./references/20-case-studies.md)


## Examples

### Example 1: Standard Scenario
Input: Design and implement a spatial computing engineer solution for a production system
Output: Requirements Analysis → Architecture Design → Implementation → Testing → Deployment → Monitoring

Key considerations for spatial-computing-engineer:
- Scalability requirements
- Performance benchmarks
- Error handling and recovery
- Security considerations

### Example 2: Edge Case
Input: Optimize existing spatial computing engineer implementation to improve performance by 40%
Output: Current State Analysis:
- Profiling results identifying bottlenecks
- Baseline metrics documented

Optimization Plan:
1. Algorithm improvement
2. Caching strategy
3. Parallelization

Expected improvement: 40-60% performance gain


## Workflow

### Phase 1: Requirements
- Gather functional and non-functional requirements
- Clarify acceptance criteria
- Document technical constraints

**Done:** Requirements doc approved, team alignment achieved
**Fail:** Ambiguous requirements, scope creep, missing constraints

### Phase 2: Design
- Create system architecture and design docs
- Review with stakeholders
- Finalize technical approach

**Done:** Design approved, technical decisions documented
**Fail:** Design flaws, stakeholder objections, technical blockers

### Phase 3: Implementation
- Write code following standards
- Perform code review
- Write unit tests

**Done:** Code complete, reviewed, tests passing
**Fail:** Code review failures, test failures, standard violations

### Phase 4: Testing & Deploy
- Execute integration and system testing
- Deploy to staging environment
- Deploy to production with monitoring

**Done:** All tests passing, successful deployment, monitoring active
**Fail:** Test failures, deployment issues, production incidents