Lakshminarayanan “Lak” Vijayaraghavan is an AR/VR systems engineer, inventor, and immersive technology architect whose career has included engineering the “Land on Mars” and “Walk on Mars” simulations at NASA Kennedy Space Center’s Astronaut Training Experience, developing foundational technologies for TikTok Effect House, engineering validation systems for Magic Leap, and creating interactive systems used across widely adopted AR and VR platforms.
Lak’s work spans immersive simulation, creator platforms, enterprise augmented reality, and AI-assisted development tools, bringing together systems architecture, real-time graphics, and human-centered design to solve complex engineering challenges.
In Part 1 and Part 2 of this series, we jumped into the conversation with Lak, including the biggest lessons he’s learned from the above endeavors, and how he approaches systems engineering and platform design today. We pick up the conversation there…
EW: Immersive computing increasingly brings together artificial intelligence, graphics, cloud infrastructure, hardware, and human-computer interaction. Which engineering disciplines do you believe will become most important as immersive platforms continue to mature?
LV: AI is already changing software engineering by accelerating many of the routine activities that consume a developer’s time. Code generation, documentation, debugging assistance, test creation, and code review can all be completed more efficiently, allowing engineers to spend more time on architecture, system design, and solving complex technical problems.
The greater opportunity, however, lies in reducing cognitive load throughout the software development process. Modern software systems involve thousands of interconnected decisions across architecture, implementation, testing, deployment, and maintenance. Engineers spend a significant amount of time locating information, understanding dependencies, switching contexts, and verifying assumptions before they can solve the actual problem. AI can help surface relevant context, summarize complex systems, identify likely impacts of a change, and automate repetitive analysis so engineers can focus their attention where it creates the greatest value.
That does not eliminate the need for engineering judgment. Architectural tradeoffs, system reliability, long-term maintainability, and product decisions still require experience and an understanding of business objectives that extends well beyond what current AI systems can infer. The engineer’s role increasingly shifts from producing every line of implementation to evaluating alternatives, validating outcomes, and making decisions that balance technical, operational, and product considerations.
I also expect AI to become more deeply integrated throughout the development environment instead of remaining a separate assistant. As these systems gain greater awareness of project structure, engineering standards, deployment history, testing results, and operational data, they will provide recommendations that are increasingly relevant to the specific context in which engineers are working.
The long-term objective is not to replace engineering expertise, but to help engineers make better decisions with less effort. The teams that combine AI with strong engineering judgment, disciplined validation, and a deep understanding of their systems will continue to build the most reliable products.
EW: AI is rapidly changing how digital experiences are created, but many discussions still focus primarily on content generation. What important engineering challenges do you believe the industry must solve before truly AI-native creator platforms become practical at scale?
LV: The first challenge is understanding intent without creating unnecessary interaction. Model quality still depends heavily on how clearly users communicate the desired result, but every additional clarification adds time, cost, and friction. The goal is not to train every creator to write better prompts. The platform should use project context, guided input, and carefully chosen questions to resolve ambiguity with as little effort from the user as possible. That also requires rethinking the interface because natural language can become an important entry point without replacing the visual controls creators need for precise iteration.
The second challenge is state awareness. Creator projects already contain scene structure, assets, scripts, previous edits, device targets, and runtime constraints. An AI system that can generate content but cannot understand that environment will make decisions using incomplete context. Practical AI-native platforms need reliable ways to inspect the current project, understand relationships between components, choose the appropriate action, and preserve continuity across multiple steps. Without that foundation, users remain responsible for coordinating the workflow themselves.
The third challenge is closing the loop through validation at a cost the platform can sustain. General-purpose models are valuable, but smaller domain-specific systems may handle many creator tasks with lower latency and more predictable behavior. Regardless of the model, the platform must still confirm that the result matches the request, fits the project, meets performance requirements, and can be deployed safely. AI-native creation becomes practical at scale only when intent, context, generation, validation, and cost are engineered as a single system.
EW: Your career has demonstrated how immersive technologies can support everything from astronaut training experiences to global creator communities. Which industries do you believe are still underestimating the transformative potential of augmented reality, and what opportunities excite you most?
LV: Education remains one of the most underutilized applications of augmented reality. Many complex subjects are still taught primarily through text, diagrams, and formulas even when the underlying concepts are spatial, dynamic, or easier to understand through direct interaction. AR can make those ideas visible within the learning environment, while AI can reduce the effort required to create immersive experiences for individual classrooms.
Consider the volume of a sphere. On paper, students may memorize that the volume is four-thirds πr³ without developing much intuition for why the value changes so dramatically. In an AR classroom, each student could manipulate a virtual sphere, change its radius, and immediately observe how the volume responds. One student could double the radius while another triples it, allowing the class to compare the results directly instead of treating the relationship as an abstract formula.
The real benefit is not simply presenting the formula in three dimensions. Students can test predictions, observe the results, compare them with one another, and discuss why the outcomes differ. Learning becomes an active process of discovery instead of memorization.
AI makes that approach much more practical because teachers should not need a specialized development team to create every lesson. A teacher could describe the concept, age group, learning objectives, and classroom activity, and the platform could generate a usable starting point that the teacher can review and refine. The underlying engineering remains within the platform through reusable interaction systems, content validation, device support, and classroom controls.
The same approach could extend to molecular structures, electric fields, orbital motion, geometry, calculus, and many other subjects where students benefit from seeing how a system changes over time. Immersive learning is especially valuable when it helps students explore relationships that are difficult to communicate through a static page.
What excites me most is the opportunity to move education from explanation alone toward guided discovery. AR provides the immersive experience, AI makes that experience easier to create, and together they can help students develop intuition about complex concepts instead of relying primarily on memorization.
EW: Every major computing platform has expanded the ways people learn, create, communicate, and experience the world around them. As someone helping shape the technologies behind immersive computing, what excites you most about where augmented reality and AI-assisted creator platforms are headed next?
LV: Two developments are especially important to me.
The first is how AI-assisted creator platforms can reduce the distance between an idea and the technical knowledge required to express it. A tremendous amount of creativity remains inaccessible because people do not know the right software, programming language, workflow, or production process. AI can absorb more of that complexity, allowing creators to begin with what they want to build instead of how every part must be implemented.
I also believe these platforms will move beyond generating content from a single prompt. They will become collaborative environments where creators can describe an idea, generate an initial version, inspect the result, make manual changes, and continue refining it with the assistance of an AI system that understands the project and its history. The creator remains in control throughout that process. The role of AI is not to replace creative judgment, but to make technical capabilities more accessible and help people iterate more quickly.
The second development is applying the same foundation to situations where better spatial understanding can improve human decision-making. The operating room is one example. Mixed reality systems can already help surgeons inspect patient-specific anatomy in three dimensions instead of relying solely on separate two-dimensional displays. As perception systems and AI continue improving, these platforms will become better at interpreting changing information and presenting the most relevant context while keeping the surgeon fully in control.
What connects these directions is the ability of AI and spatial computing to make complex systems easier to understand and interact with. In a creator platform, that means helping someone transform intent into a working experience. In a medical environment, it means helping someone interpret complex spatial information with greater confidence. Building systems that make advanced capabilities more understandable, controllable, and useful is what excites me most about the future of immersive computing.
Ellen F. Warren
AR Insider Guest Author