Corbin Klett was standing on a runway in Houston in an airplane factory troubleshooting a wiring problem in a jet prototype when he had a realization.
It was his third wiring problem this week, and it was only Tuesday. The problems he was fixing should have been solved upstream in design.
Despite working with cutting-edge aerospace technology, he was still managing electrical systems the same way as engineers from the 70s —with scattered documentation, manual processes, and frequent clerical errors that could ground test flights.
So Klett talked coworker and long-time friend Antony Samuel into building Artifact, what they describe as "the VS Code for hardware engineers."
Three months later, they got their acceptance into YC, and they set off to build a comprehensive design environment including AI copilots to automate the complex, error-prone work of electrical system design.
Tackling the 2D Problem Space
What makes Artifact particularly interesting is their focus on what Klett calls the "2D problem space"—using AI to generate structured 2D drawings like electrical schematics, rather than the 1D text editing that dominates current AI applications or the 3D CAD modeling that other companies are attempting.
Schematics are made of discrete building blocks that need to be connected logically — which means that LLMs can generate them in a structured output format the way they’d generate a block of JSON.
Preview of Schematic Design in Artifact.
Their initial users are small to mid-sized aerospace companies using the platform to design electrical wiring harnesses.
"Right now there's a lot of manual work that goes into making these drawings,” Klett says. ”[But with] the copilot, you tell it what you want, it produces the drawing, saves you time, increases reliability."
Their moonshot vision is to create a “Jarvis”-type system allowing individual engineers to design full systems without being bogged down by bureaucracy.
Building on Mastra + Next.js
Klett and Samuel started building their agentic features with Langchain, but found it messy and hectic. That’s when they discovered Mastra.
Their architecture combines Mastra’s workflow orchestration to manage tool use (in their case, for schematics generation) with Assistant UI and Next.js to create a full-stack, Typescript agentic aerospace app. They’re deploying on AWS EC2 clusters.
Looking Forward
With the growing focus on defense and industrial tech, Klett and Samuel were able to raise seed funding from Mike Maples at Floodgate.
Artifact's timing couldn't be better. "Now you've got all these new hard tech companies and we need a revolution in that space," Klett laughs.
And so far, they have a concrete place to start, by hardware engineers actually be engineers, rather than documentation managers.