Audrey Woods | MIT CSAIL Alliances
An estimated 20% of every dollar spent on manufacturing is wasted, totaling up to $8 trillion a year, more than the entire annual budget for the U.S. federal government. While industries like healthcare and finance have been rapidly transformed by digital technologies, manufacturing has relied on traditional processes that lead to costly errors, product delays, and an inefficient use of engineers’ time. Armed with emerging AI technology and a big problem to solve, three CSAIL scientists set out to change this.
Foundation EGI—conceived in the lab of MIT CSAIL Professor Wojciech Matusik and co-founded by former MIT CSAIL researchers Mok Oh, PhD ‘02, and Michael Foshey, MEng ‘23—has launched the first domain-specific agentic AI for engineering. Their platform was specifically created to “transform physical engineering by codifying the entire process into a Domain-Specific Language (DSL) to bring the power of software engineering to hardware design for the first time.” Accelerating every stage of the product development cycle, Foundation EGI aims to automate workflows, save time, unify engineering knowledge, and foster innovation. Their product promises traceability, auditability, and a natural language interface based on a real-world intelligence which eliminates the risk of hallucinations.
The company was originally launched in 2023, but has been operating in stealth mode as the team quietly perfected their technology and assembled an expert team. Two of the co-founders, Professor Matusik and Foshey, participated in the 2024 paper “Large Language Models for Design and Manufacturing,” which laid out the academic basis of their idea and crystallized the mission of the company. Now, Foundation EGI is backed by an over-subscribed $7.6M seed round which includes MIT’s E14 Fund, Union Lab Ventures, Stata Venture Partners, Samsung Next, GRIDS Capital, and Henry Ford III. Their product is being beta tested at leading Fortune 500 companies and is already showing measurable results in fewer production errors, stronger revenue outcomes, and a faster design-to-market cycle.
For Mok Oh, Co-founder and CEO, “there’s a perfect storm brewing with the potential to dramatically transform design and manufacturing industries.” He clarifies that while there’s widespread excitement about AI models like ChatGPT, “the broad, generic open-source AI technologies are not a good match for narrow or domain-specific use cases; they are often unreliable and prone to hallucinations.” However Foundation EGI’s agent, a domain-specific LLM fine-tuned with engineering data and processes, was designed with manufacturing tasks in mind. It’s well-versed in the particular tools and vocabulary of engineering, but also offers the approachable natural language interface of widely-used ChatGPT-style models.
The Foundation EGI team aspires to not just launch a product but a whole new subset of AI technology: Engineering General Intelligence. Professor Matusik, who has built his career applying computer science tools to manufacturing and design, says, “EGI will unleash the creative might of a new generation of engineers. Expect leaps and bounds in agility, innovation and problem-solving.” By cutting costs, dramatically increasing efficiency, and streamlining product development, EGI has the potential to revolutionize manufacturing and every industry it touches.
Learn more about Foundation EGI on their website, see a video of their platform in action here, and read about other CSAIL Spinouts on our case study page.
