Lu Yang argues that mechanical engineers are working with tools twenty years out of date. Her company Luminvera seeks to fix that.
The disconnect inside modern industrial engineering is rarely the topic of conferences on the future of artificial intelligence. The conferences talk about digital twins, AI-driven automation, and the manufacturing layer that will support the next generation of humanoid robots. The engineers who actually design those systems spend most of their working hours in front of two-dimensional screens, translating spec sheets that arrive as plain text into computer-aided-design models that exist on a flat plane that the eventual physical machine will never occupy.
The phrase Lu Yang uses for that condition is the “2D Stone Age.” She is the founder and chief executive of Luminvera Inc., a Silicon Valley company building augmented-reality solutions with enterprise artificial-intelligence features for industrial engineering teams, and she has spent the better part of ten years inside the engineering organizations the phrase describes.
Luminvera was incorporated in March 2026. The thesis the company was founded around, framed in the language Lu Yang has been refining for several years, is that the most expensive engineering talent in the world is being applied to the wrong problem in the wrong dimensional register. The pivot the company is trying to drive in the category, again framed in her terminology, is the move from “analog management” to what she calls “immersive engineering.”
Lu Yang’s background is the part of the company that explains the rest of it. She trained originally as an IT project manager. She spent the substantive portion of her career at Bosch and Mercedes-Benz, two of the largest and most engineering-conservative manufacturers in the world, leading digital-transformation work across more than three thousand engineering units. She built and rolled out AI-specific quality methodologies and regulatory-training programs across an entire business division. She led interdisciplinary product teams across at least four international locations. The work was the work of taking ideas from the consumer-software conversation and making them survive contact with engineering-floor realities.
Founders in this category have historically had software backgrounds. The distinction matters more than it appears to. A software founder building an AR product for industrial use has to learn the domain through customers. An industrial-engineering founder building the same product has the failure modes already cataloged. The difference shows up in what the product is willing to do and refuse to do, and Lu Yang has built Luminvera around the conviction that an XR immersive platform for an industrial engineer is a different product than an AR wearable for a knowledge worker.
The category Luminvera is targeting has been defined inside the industry, in the terminology Lu Yang uses, as the Context Gap. The Context Gap is the distance between the two-dimensional CAD model an engineer designs against and the three-dimensional physical machine the model will eventually become. The gap is filled, in current practice, by manual translation, intuition, and a long chain of paper and email handoffs. The cost of bridging the gap, expressed in engineering hours, is large enough that it would dominate the cost line of any manufacturer that tracked it carefully. Most do not track it carefully.
Lu Yang has cataloged the failure modes inside the Context Gap in terminology she uses publicly. The most foundational, in her framing, is what she calls “Dimensional Dissonance,” the loss of spatial reasoning that occurs when complex three-dimensional systems are simulated through the keyhole of a flat two-dimensional screen. Three additional failure modes operate alongside Dimensional Dissonance, including the manual translation of multi-thousand-page customer specifications, the persistence of paper-based collaboration between engineering centers and manufacturing plants, and the gating role of compliance documentation authored in static text and two-dimensional images for systems that exist physically in three dimensions. The design of Luminvera’s product is organized against the four of them.
Each failure mode appears, in the current state of the industry, as a productivity problem. Lu Yang’s argument is that the productivity framing understates what is happening. Engineers operating in the conditions she has described are working under structural constraints the industry has accepted as immutable, and the constraints are, in her view, not immutable.
What Luminvera builds against that argument is a wearable AR device with an enterprise AI layer. The hardware is designed for the factory floor. The software layer is designed against the regulatory and governance constraints that operate in heavily regulated markets, particularly in Europe, where her Bosch and Mercedes-Benz years gave her direct exposure to what compliance requires. The product’s positioning is hands-free mechanical operations and real-time design adjustments, with the AR layer carrying the spatial context that the desktop screen flattens out and the AI layer carrying the unstructured-specification translation that the engineer would otherwise be doing manually.
The credibility of the design choices comes from the credibility of the founder. Lu Yang sits at the intersection of three disciplines that do not normally appear in a single career: industrial engineering, AI governance, and immersive technology. The credibility of an AR product targeting industrial workflows comes from whoever can speak to all three. Most founders in the category cannot. Lu Yang can, and the early Luminvera roadmap reflects it.
The conversation about whether industrial engineers should be using AR and AI tools differently from how they currently are is one the manufacturing industry has been having quietly for several years. The conversation has produced a small number of pilot deployments, mostly inside corporate innovation programs that have not survived contact with line operations. What it has not produced, until now, is a company built end-to-end around the Context Gap as the primary problem. Luminvera is the first.
Whether the company succeeds in pushing the category from analog management toward immersive engineering, in the language Lu Yang uses, is the question the next several quarters will answer. The thesis is built. The hardware is in development. The first customers are in conversation. Lu Yang has been preparing for the conversation since well before Luminvera existed.
