Is Spatial Computing Developer Worth Learning in 2026? AR/VR Career Path, Salary, and the Vision Pro and Android XR Wave

A lot of people asking whether spatial computing is worth learning are not really asking about technology. They are asking whether they are about to waste a year of their career on another XR hype cycle. That is the right question.

My answer is yes: spatial computing is worth learning in 2026, but only if you understand that “AR/VR developer” is not one job, that the consumer headset market and the enterprise market are behaving very differently, and that the center of gravity has shifted. In 2025, the XR market shipped about 14.5 million devices, up 41.6% year over year, but that growth did not come from standalone VR headsets suddenly roaring back. It came from smart glasses, which represented roughly half of all XR shipments, while standalone VR and mixed reality headsets contracted in the first half of the year before recovering later. Reality Atlas also reports that more than 75% of Fortune 500 companies are actively using XR technologies across 2025–2026.

That is why the old “is VR dead?” framing misses what is actually happening. Standalone VR is no longer the whole story. Apple pushed spatial computing back into serious executive conversations with Vision Pro, even though its unit sales stayed niche. Treeview’s 2026 market report estimates Vision Pro at roughly 475,000 lifetime units by the end of 2025, with 390,000 units in its launch year and 85,000 in 2025, and it pegs enterprise buyers at 75% of purchases with 50+ Fortune 100 organizations using the device. Samsung then launched Galaxy XR in October 2025 as the first shipping Android XR headset at $1,799, giving Google its first credible operating-system-level entry into this market outside of phones.

If you are choosing a career path, the implication is simple: spatial computing is real enough to build a career on, but not broad enough yet to reward vague learning. The people who win in this market are not the ones who “learn XR” in the abstract. They pick the right layer of the stack, build a portfolio around shipping constraints, and move upward from graphics skills into interaction, then from interaction into perception and spatial AI. That distinction is what most blog posts miss, and it is the distinction that actually determines whether this field feels like a niche curiosity or a sensible long-term specialization.

Why Smart Glasses Just Outsold VR Headsets and Why “AR/VR Developer” Still Means Four Different Jobs

The biggest mistake newcomers make is assuming the market question and the career question are the same question. They are related, but they are not the same.

The market question is whether spatial computing has finally crossed from demos into durable use. The career question is whether companies now need enough developers with the right skills to justify your learning time. Those answers are not identical. A platform can have underwhelming consumer sales and still create excellent developer opportunities if enterprises are buying it, if platform vendors are investing in tools, or if cross-platform skills transfer cleanly into adjacent work. That is exactly what Treeview’s 2026 XR market report shows across Vision Pro, Quest, and Android XR.

The shift in hardware categories matters because it changes what companies hire for. When smart glasses become the volume driver, teams need more than immersive-scene developers. They need people who can work with multimodal inputs, context-aware interfaces, lightweight 3D assets, perception APIs, and world-anchored UI patterns. Android XR’s own developer documentation is explicit that apps can be spatialized from existing Android experiences, that developers can add spatial panels, 3D models, and immersive environments, and that ARCore for Jetpack XR brings perception features such as anchors and semantic segmentation. That is a different hiring profile from “make a room-scale game in Unity.”

Apple’s side of the market is different but points in the same direction. According to Apple’s visionOS developer documentation, apps in Shared Space automatically benefit from ARKit, and when apps move into a Full Space they can use ARKit capabilities such as world mapping, segmentation, environment lighting, and hand-related sensing. Apple has also expanded business-focused APIs, including enterprise camera access, which is the kind of platform move you make when you believe the near-term revenue story is professional workflow, simulation, training, and field operations, not App Store-scale consumer volume.

Meta remains the clearest proof that there is still a practical market for headset development, but even there the skill profile has matured. Meta’s developer stack now leans heavily into natural input, shared spatial anchors, body and eye tracking on supported hardware, haptics, and spatial audio, not just controller-driven room-scale interactions. Hand tracking, Meta says plainly, is used to deliver more natural interactions and a stronger sense of presence, while Shared Spatial Anchors create a shared world-locked frame of reference for colocated users. Those are the ingredients of spatial interaction engineering, not just general app development.

This is why the title “AR/VR developer” has become sloppy to the point of being misleading. In practice, employers lump together at least four different roles under that label. Some want a graphics-heavy Unity or Unreal generalist. Some want a platform-specific app developer for Quest, visionOS, or Android XR. Some want a senior interaction engineer who understands hands, eyes, haptics, audio, and collaborative spatial systems. Increasingly, some want a hybrid who can connect perception, scene understanding, or generative AI to real-time spatial interfaces. That is a huge spread in difficulty, compensation, and promotion path.

I have hired developers who were excellent at shaders and terrible at interaction design. I have also hired developers who could build slick demos but had no feel for frame budget, input ambiguity, or motion comfort. In spatial computing, both kinds stall. The market is mature enough now that “I know Unity” is no longer a useful self-description. The question that matters is: which layer of the work can you ship reliably?

The Spatial Computing Stack

To make that distinction concrete, I use a framework I call the Spatial Computing Stack. It is the clearest mental model I know for understanding why salary ranges, job descriptions, and career outcomes vary so wildly under the same “AR/VR developer” label.

The framework has four layers. They are cumulative rather than mutually exclusive. You do not throw away Layer 1 when you move into Layer 3. You build on it. And that matters, because the fastest promotions I have seen in spatial teams came from developers who treated each layer as a durable competence, not a temporary stepping stone.

Layer 1: 3D and graphics foundation. This is where most people start, whether they realize it or not. The output here is not “an XR career.” It is general spatial literacy: Unity or Unreal fundamentals, scene graphs, coordinate systems, 3D math, materials, lighting, performance instincts, shader basics, input loops, build pipelines, and often a first pass at WebXR. MDN’s WebXR documentation is a useful marker of this layer because it focuses on the mechanics of device pose, stereoscopic rendering, input sources, anchors, depth sensing, and hit testing. If you cannot reason comfortably about spaces, transforms, and render cost, every later layer gets harder.

Layer 2: XR app developer. This is the layer most job posts actually mean. The output is a working application on a target platform. On Apple, that means understanding visionOS windows, volumes, spaces, SwiftUI or RealityKit, and ARKit-driven blended content. On Meta, that usually means Unity or Unreal plus the relevant Meta XR SDKs, Interaction SDK, anchors, input, store constraints, and performance tuning. On Android XR, it can mean Jetpack XR, SceneCore, ARCore for Jetpack XR, or OpenXR game-engine workflows with Unity, Unreal, or Godot. The defining trait of Layer 2 is not that you know “XR.” It is that you can ship on an actual device family with platform-specific constraints.

Layer 3: Spatial interaction engineer. This is where generic XR teams separate into seniors and specialists. The output here is not merely that the app works; it is that the app feels native to spatial computing. That means hand tracking that does not produce accidental activations, gaze interactions that do not create fatigue, haptics that reinforce rather than distract, audio that supports presence and orientation, and shared spatial experiences that keep alignment and state coherent across multiple users. Android XR’s design guidance explicitly calls for multimodal inputs including hand and eye tracking, voice, controllers, mice, and keyboards; Apple’s visionOS stack supports spatial SharePlay and shared contexts; Meta’s stack covers hand tracking, haptics, spatial audio, body and eye tracking, and shared spatial anchors. This layer is where interaction design becomes engineering, not decoration.

Layer 4: Spatial AI engineer. This is the newest rung, and it is the rung a lot of people talk about before they are ready for it. The output here is a spatial application that uses perception or AI as a first-class part of the workflow: object tracking, scene meshing, semantic segmentation, depth, contextual assistance, world-aware search, or AI-assisted 3D content generation. Apple’s visionOS stack now includes object tracking and enterprise camera access for specialized business workflows; Android XR’s OpenXR and ARCore stack exposes scene meshing, depth textures, hand and eye tracking, passthrough, and perception APIs; Google’s Android XR glasses demos are explicit that Gemini can “see and hear what you do” to understand context. This is where spatial computing starts to collide with computer vision and multimodal AI without becoming the same job as either one.

The table below is the practical version of the framework. The pay figures are not perfect, because salary aggregators split the market by title rather than by true skill layer. But that title mismatch is part of the point: the field pays differently because the field is really several fields wearing one label.

The salary signals in that table come from current U.S. aggregator pages for Glassdoor’s Unity3D (AR/VR) Developer and AR/VR Developer listings, plus ZipRecruiter’s AR/VR Developer, AR/VR Engineer, and VR Software Engineer pages. I would not treat them as exact rates for your future offer. I would treat them as directional evidence that title granularity matters, and that more specialized interaction and engineering-heavy roles tend to clear more money than broad “developer” labeling.

The realistic on-ramp between layers is also predictable.

From Layer 1 to Layer 2. Stop building engine demos for other developers and start building product slices for actual users. That means onboarding, quitting and resuming sessions, handling permissions, explaining hand interactions, running on battery-constrained hardware, and packaging for a real platform. A gorgeous shader toy does not make you an XR app developer.

From Layer 2 to Layer 3. Get obsessed with interactions that fail in the real world. In code review, I trust the developer who worries about reach distance, accidental pinches, tracking loss, re-centering, shared state drift, and comfort. That person is thinking like a spatial interaction engineer.

From Layer 3 to Layer 4. Learn where perception actually creates value. Not every app needs object tracking. Not every workflow needs a camera feed. What separates good Layer 4 engineers is not AI enthusiasm; it is judgment about when real-time scene understanding or AI-generated content materially improves the experience.

What a Spatial Computing Developer Actually Does in a Normal Week

One reason people misjudge this field is that their mental picture of the job is wrong. They imagine either a game developer in a headset lab or a futuristic AI researcher. Most working weeks in spatial computing look much less glamorous and much more product-oriented.

A normal Layer 1 week. You spend time building and breaking scenes. You import assets, fix pivots, optimize textures, adjust materials, debug transform hierarchies, test controller or hand input, and spend an annoying amount of time dealing with lighting, occlusion, and frame timing. If you touch WebXR, you are also thinking about browser support, secure contexts, and whether your experience degrades gracefully outside fully immersive devices. This is the layer where you learn the fundamentals that people coming from traditional web development often underestimate. If you want a clean contrast, read how the frontend web developer career path differs from 3D and WebXR work. The web stack can get you into the conversation, but WebXR still asks you to think in terms of spaces, poses, stereoscopic rendering, and device-driven input rather than 2D layouts and DOM events.

The hiring bar at this layer is not “knows Unity” anymore. It is “can show me three projects where the scene is stable, the interaction loop is coherent, and the frame rate does not collapse.” On promotion, Layer 1 developers typically prove that they can own a feature end to end without needing someone else to constantly rescue the build, the render pipeline, or the interaction logic.

A normal Layer 2 week. You are much closer to shipping than to experimenting. On visionOS, that can mean reconciling what belongs in a window versus a volume versus a Full Space, working with RealityKit, integrating SwiftUI with 3D content, and making sure your app behaves correctly around ARKit permissions and scene blending. On Android XR, it might mean adapting a large-screen Android app into spatial panels, adding SceneCore entities, wiring up ARCore for anchors or segmentation, or using game-engine support through OpenXR. On Meta Quest, it often means input testing, store readiness, comfort checks, anchor persistence, controller and hand parity, and performance profiling across target devices. Apple, Google, and Meta all frame this work as building with familiar tools extended into XR, but “familiar” does not mean trivial. It means you get leverage from what you already know, not a free pass around device constraints.

This is the layer where portfolios improve the fastest when developers stop trying to impress and start trying to reduce risk. A hiring manager does not need your moonshot metaverse vision. They need to know whether you can ship a narrow but credible spatial app that survives real device behavior. In practice, that usually means a training flow, a design-review tool, a collaborative viewer, a guided workflow, or a productivity slice, not a giant open-world XR platform.

A normal Layer 3 week. The calendar becomes more multidisciplinary. You are in design reviews, usability sessions, and often platform-specific experimentation because you are no longer just implementing features; you are deciding how people should do things in 3D space. Android XR’s guidance emphasizes multimodal inputs and warns against relying on a single method. Apple’s visionOS work around SharePlay and shared spatial experiences demands synchronization, state clarity, and social presence. Meta’s tools for haptics, spatial audio, body tracking, and shared anchors turn interaction quality into a systems problem. That is why this role often sits at the boundary between engineering and product design. If you want a separate view of the design side, see the UI/UX design career path, salary, and portfolio expectations. Spatial interaction engineers are not replacing UX designers; they are the people who turn interaction intent into device-native, trackable, comfortable behavior.

The developers I have seen move fastest into Layer 3 all shared one habit: they treated ambiguity as part of the job. They did not wait for a designer to specify the exact pinch gesture, or for QA to discover that gaze dwell was fatiguing, or for users to report that shared alignment drifted in multiplayer. They anticipated those failures. That is seniority in spatial computing.

A normal Layer 4 week. Now the work starts looking like productized perception. You might prototype object tracking for a maintenance workflow on Vision Pro, use enterprise camera access for a specialized in-house tool, test depth or scene meshing on Android XR, or connect Gemini’s contextual assistance to a spatial experience that understands what the user is looking at. On Android XR, official OpenXR support includes passthrough, scene meshing, depth textures, hand and eye tracking, and performance features such as eye-tracked foveation; Google’s glasses demos frame Gemini as an assistant that can see and hear context. On Apple’s side, enterprise APIs expose main-camera-driven workflows, and object tracking continues to expand for precision use cases.

This is also where role confusion becomes expensive. A spatial AI engineer is not the same as a computer vision engineer. The closest working definition is this: the computer vision engineer is usually building or adapting the perception model, while the spatial AI engineer is building the application layer that consumes perception, contextual AI, or generated 3D content inside a real-time spatial interface. If you want the deeper distinction, read how computer vision engineering differs from building the applications that use those models. And if you want macro context for why generative systems are becoming part of this discussion, see the broader trends shaping generative AI models. The overlap is real, but the jobs are not interchangeable.

The promotion logic in Layer 4 is also different. You do not get there by sprinkling AI into a demo. You get there when leaders trust you with privacy-sensitive inputs, real-world sensing, model-serving constraints, and UI decisions that translate noisy perception into reliable user benefit. In other words, Layer 4 is less about AI novelty than about judgment.

Is This Wave Real or Another Hype Cycle?

This is the part most articles dodge. I will not.

The honest answer: the 2026 spatial-computing wave is real, but the form it is taking is not the one early metaverse hype promised.

If by “real” you mean “hundreds of millions of consumers are buying premium headsets and spending all day in immersive 3D operating systems,” then no. That world is still not here. Treeview’s 2026 XR market report estimates Vision Pro at roughly 475,000 cumulative units by the end of 2025 and Samsung Galaxy XR at roughly 40,000 units in 2025. Those figures are far too small to support a broad consumer gold rush; they are clearly early-adopter numbers, not mass-market device volumes.

But if by “real” you mean “the hardware, operating systems, developer tools, and enterprise use cases have finally matured enough to support durable hiring and serious product work,” then yes, much more than in past cycles. Apple has leaned into business use cases publicly. It said in 2024 that more than half of Fortune 100 companies had purchased Vision Pro, and Apple’s business messaging for Vision Pro centers on design collaboration, employee training, and guided fieldwork. Treeview’s report goes further and estimates that enterprise buyers account for 75% of Vision Pro purchases. That is not consumer breakout. It is something more practical: enterprise validation.

Android XR is also more serious than many developers first assumed. Google positions it as an open, unified platform for headsets and glasses, and Samsung calls Galaxy XR the first product in a broader Android XR ecosystem spanning headsets and AI glasses. Google’s 2026 updates expanded engine support across Unity, Unreal, and Godot, emphasized OpenXR workflows, and pushed the ecosystem beyond a single Samsung headset with XREAL Aura entering the pipeline. That is the behavior of a platform company trying to build a category, not of a company running a short-lived experiment.

The strongest evidence that this is a real wave, however, is not Vision Pro or Galaxy XR in isolation. It is the hardware mix shift. When smart glasses become roughly half of all XR shipments, and the total market moves to around 14.5 million devices with smart glasses driving most of the growth, the lesson is not “VR finally won.” The lesson is that the market found a more wearable, more socially acceptable, and more everyday-adjacent form factor than bulky headsets. That is a structural change in what people are actually willing to buy and use.

There is still a hype risk, though, and you should factor it into career decisions.

What remains overhyped. The idea that every developer should rush into building for premium consumer headsets is still overhyped. The install base is too fragmented, the content economics are too uneven, and platform priorities are still shifting. Apple is highly influential but still niche in units. Samsung and Google are promising but early. Treeview’s 2026 market report shows smart-glasses volumes growing fast, but the best-selling smart-glasses products have so far leaned more toward AI assistance and media capture than toward rich third-party app ecosystems. That means the safest career bet is not “pick one device and hope.” It is “build skills that transfer across the stack.”

What is no longer hype. Enterprise workflow value. Multimodal interaction. Cross-device operating systems. Spatial development inside familiar tools. AI context in wearables. Those are not speculative anymore. Android XR explicitly supports existing 2D Android apps, spatialized app adaptations, and full immersive experiences with Jetpack XR or engines. Apple continues to deepen visionOS for productivity and enterprise use. Meta continues to supply the most mature mainstream headset development environment. The ecosystem is still messy, but it is materially less fictional than the metaverse era pitch deck version of XR.

My own read is this: 2026 is the year spatial computing stopped being one story. There is now a consumer-wearables story, an enterprise-spatial-workflow story, a headset-gaming-and-training story, and an emerging spatial-AI story. That fragmentation is not bad news for developers. It is exactly what makes the field worth learning. Specialized markets are where technical careers become more durable.

Salary and How to Break In

Let’s talk money without pretending the data is cleaner than it is.

Glassdoor’s live U.S. AR/VR Developer page currently shows an average salary of about $114,129 per year and roughly $55 per hour, with top earners around $177,838 and a typical range of $90,944 to $144,509. ZipRecruiter’s live AR/VR Developer page lists $109,905 per year, with the 25th to 75th percentile band at $84,000 to $134,500. In plain English, the market signal is that a general AR/VR developer in the United States sits a little above the low six figures on average, with a wide enough spread that specialization, geography, and title matter a lot.

The first thing I would tell any newcomer is not to obsess over salary averages before understanding title inflation. The same broad skill family gets labeled as AR/VR Developer, Unity3D AR/VR Developer, VR Developer, AR/VR Engineer, XR Engineer, Spatial Software Engineer, and platform-specific roles tied to Quest or visionOS. Those pages are useful directional signals, but they are not a precise market map.

Even so, they do hint at progression. Glassdoor’s Unity3D (AR/VR) Developer page sits lower than the broader AR/VR Developer page, at roughly $103,543 average. ZipRecruiter’s current AR/VR Engineer page shows roughly $115,864 average, and its VR Software Engineer page rises further to about $147,524 average. I would not describe that as a guaranteed linear ladder; reality is messier than that, but I would absolutely describe it as evidence that engineering-heavy, systems-heavy, and more senior titles tend to command better compensation than entry-level engine generalism.

So which layer should you target first?

If you are coming from game development or graphics programming, start at Layer 1 and move to Layer 2 quickly. You already have the hardest transferable advantage: comfort with 3D systems, engine tooling, performance, and iteration loops. Your risk is getting trapped in demo-land. The fix is to build two or three product-shaped portfolio pieces for actual devices. One Quest piece, one platform-native or cross-platform spatial app, and one small interaction-heavy prototype is a much better portfolio than ten disconnected visual experiments.

If you are coming from mobile development, target Layer 2 directly. visionOS and Android XR are the most natural bridges because both explicitly extend existing app-development models. Android XR says you can spatialize existing Android apps and use familiar Android tools. Apple emphasizes familiar frameworks with spatial extensions. If you are already strong in native app architecture, state management, and shipping discipline, your real gap is not “learning XR.” It is learning how spatial interaction changes layout, input, comfort, and world-awareness.

If you are coming from frontend or design-engineering work, be careful not to mistake 3D UI for 2D UI in a different canvas. This is the most common false start I see. Floating panels are not enough. Spatial apps live or die on depth, reach, targeting precision, feedback, comfort, and environmental context. That is why WebXR can be a useful bridge but not a substitute for engine or device experience. It teaches the right mental model, but the hiring market still rewards people who can prove they have touched real device constraints.

If you are coming from computer vision or AI, resist the temptation to start at Layer 4. Most of the strongest spatial AI hires I have seen were first credible Layer 2 or 3 developers. They already understood device permissions, real-time UX, and interaction friction, so when they added perception or AI they produced features people could actually use. Pure model fluency without interaction fluency often produces demos that feel magical in a pitch and useless in a product.

The tactical roadmap I recommend in 2026 is more boring than people expect, but it works.

First, build one strong Layer 1 foundation project: a performant 3D prototype with clean interactions, asset hygiene, and clear technical decisions. If you cannot explain your scene graph, object lifecycles, and performance tradeoffs, you are not ready to move on.

Second, build one narrow Layer 2 shipping project on a real device family. Not “coming soon.” Not simulator-only. Real hardware. The project should solve a clear use case, even if small: a spatial product viewer, a room-anchored instruction flow, a collaborative tabletop demo, a design-review experience, or a simple workflow assist tool.

Third, build one interaction-first project that forces you into Layer 3 thinking. Hand-only controls. Shared anchors. Spatial audio cues. Gaze-assisted far interaction. Haptics tied to object state. The point is to prove that you understand that spatial UX is a systems problem.

Only after that would I build a Layer 4 prototype. And when you do, keep it ruthlessly grounded. Object tracking for a training workflow. Gemini-assisted contextual guidance in a spatial app. Scene meshing for world-aware placement. AI-generated 3D content with a clear user outcome. Not “AI but in 3D.” That phrase should terrify you.

Portfolio reviewers in this field scan for one thing above everything else: whether you understand that devices are the product. A good spatial computing engineer does not build abstract experiences. They build experiences that work on real hardware, with real sensor behavior, real input ambiguity, real comfort limits, and real platform rules. That is why the best break-in strategy is still practical shipping, not speculative branding.

FAQ

Is AR/VR developer a good career in 2026? Yes, with an important qualifier. It is a good career if you approach it as a specialization inside software engineering, graphics, interaction, or applied AI, not as a vague bet on “the metaverse.” Salary data remains solidly above the low six figures in U.S. aggregators for general AR/VR developer titles, and the market evidence now includes enterprise adoption, platform investment from Apple and Google, and sustained device development from Meta. What is not sensible is learning this field as if consumer headset scale has already arrived. It has not.

Do I need a computer vision background to work in spatial computing? No. Most spatial computing jobs are still Layer 1 through Layer 3 jobs: graphics, app shipping, and interaction engineering. You usually only need strong computer-vision depth when you move into spatial AI or perception-heavy workflows, such as those built on Android XR’s OpenXR scene-meshing and depth capabilities. Even then, many teams split the labor so that CV specialists build or adapt the perception stack while spatial developers build the app layer that consumes it.

Should I learn visionOS, Android XR, Quest, or all three? If you want the safest first platform for hands-on learning, Meta Quest remains the friendliest practical playground because its tooling is mature and the interaction surface is well understood. If you have strong Apple-native experience, visionOS is the cleanest way to turn that into spatial development. If you want to position for the next wave of headsets and glasses, Android XR deserves serious attention because Google is building it as a unified platform for both categories and expanding support across Unity, Unreal, Godot, OpenXR, and spatialized Android apps. I usually recommend learning one shipping platform deeply and understanding the others well enough to port your mental model.

Is the metaverse dead? As a dominant product vision, mostly yes. As a bundle of underlying skills, including real-time 3D, spatial interfaces, world-aware computing, shared spatial presence, and multimodal interaction, absolutely not. The industry has moved from a grand unifying fantasy to several more practical markets, a shift reflected in Treeview’s 2026 XR market report: enterprise XR, smart glasses, premium mixed reality, collaborative spatial tools, and emerging spatial AI. That is healthier for careers than the old “one virtual world to rule them all” story ever was.

How long does it take to become a spatial computing developer? If you already know software development, you can reach a credible Layer 2 portfolio in a few focused months of serious work. If you are starting from scratch, expect longer because Layer 1 fundamentals matter more than people think. The trap is chasing platform APIs before you understand spatial thinking. In this field, weak fundamentals show immediately.

Is spatial computing mostly enterprise, then? Right now, much more than the loudest consumer marketing suggests. The Treeview/Reality Atlas 2026 market report estimates that enterprise buyers account for 75% of Vision Pro purchases, that more than 50 Fortune 100 organizations use the device, and that over 75% of Fortune 500 companies are actively using XR across 2025–2026. That does not mean consumer demand is irrelevant. It means enterprise is currently the cleaner business case.

What is the best background for moving into spatial computing? Graphics programming and game development are still the strongest raw feeders because they create intuition for 3D systems, performance, and interaction loops. Mobile development is a close second now that visionOS and Android XR both extend familiar app-development models. Design engineers can transition well if they respect the technical realities of 3D interaction. Pure web developers can bridge in through WebXR, but they usually need more time on device-native constraints before they look strong in interviews.

What actually makes someone promotable in this field? Ownership under device constraints. Juniors can usually get a demo working. Mid-level developers can make it ship. Seniors can make it feel native, stay performant, survive edge cases, and support real users. The jump from “XR developer” to “spatial interaction engineer” happens when people trust you with ambiguity, input design, performance tradeoffs, shared experiences, and product judgment.

If I only have time to learn one thing first, what should it be? Real-time 3D fundamentals. Not because they are glamorous, but because they compound. The strongest interaction engineers and spatial AI specialists I know still fall back on the same Layer 1 instincts: transforms, timing, rendering, asset cost, state flow, and comfort. In spatial computing, fundamentals are a career advantage, not a beginner phase.