MIT’s JARVIS Challenge: Can AI Design a Jet Engine?

The Dawn of AI in Tough-Tech Engineering

At a glance, Artificial intelligence, particularly generative AI and large language models (LLMs), has dramatically reshaped software engineering, automating code generation and vulnerability detection. But how does AI fare when the challenge involves conceiving, designing, and fabricating complex physical systems, such as a jet engine? This was the central question behind the JARVIS Challenge (Jet-engine AI Research and Validation Intensive Sprint) at MIT.

Meanwhile, The JARVIS Challenge aimed to determine if AI could significantly compress the traditional design-build-test cycle for safety-critical hardware. Over a single semester, MIT undergraduates were tasked with exploring AI’s potential to accelerate and improve their engineering endeavors.

“The JARVIS challenge showed that AI can substantially accelerate safety-critical hardware engineering, but engineering judgment remains the decisive differentiator. An AI-native engineer is not defined by using AI, but by leading it — knowing when to trust it, when to challenge it, and how to translate AI outputs into working hardware.” – Professor Zolti Spakovszky, Director of the MIT Gas Turbine Laboratory.

The Challenge: Building a Jet Engine in Weeks

Students were given a mere four weeks to design, fabricate, assemble, and test a small gas turbine aero engine. Their objective: construct a “JARVIS-class” single-spool jet engine capable of producing 50–100 pounds of thrust, running on Jet-A fuel, and completing five 60-second runs. Teams had complete autonomy over design, materials, and fabrication choices.

In practical terms, Thirty-one students, representing various engineering departments, formed seven teams. Many participants had limited prior experience in turbomachinery, compressible flows, or even thermodynamics. They were literally learning to build a gas turbine from the ground up.

Resources at Their Disposal:

• Access to MIT’s machine shops and manufacturing vendors.
• Commercial software like Concepts NREC, SolidWorks, and ABAQUS.
• Test rigs for component characterization and assembly.
• Unlimited use of MIT Parley, a new platform aggregating frontier LLMs through a single interface, allowing challenge leaders to monitor AI usage.

Sponsors like MIT Lincoln Laboratory, the Department of Mechanical Engineering, Safran, Voyager Technologies, and Beehive Industries supported the initiative, driven by both recruitment interest and genuine curiosity about AI’s impact on engineering workflows.

AI as an Engineering Copilot: Strengths and Stumbles

For example, In the initial week, teams leveraged AI for various tasks, including summarizing textbooks, learning design software, sourcing vendors, creating spreadsheets, answering specific questions, and performing comparative analyses. One team even created an AI agent in Parley to serve as their project manager.

However, as teams moved into detailed CAD designs and prototyping, AI’s limitations became apparent. While LLMs like Claude and ChatGPT were useful for offering design alternatives and filling knowledge gaps, their propensity for hallucinations, sycophancy, and a fundamental lack of physical understanding began to undermine student confidence and slow progress.

“AI is a helpful tool, great at finding information, helping organize things, and can write well, but it can’t do design. The moment the engineer doesn’t know what is going on and the AI is in charge is the moment the design becomes unreliable, at least with AI at its present capabilities.” – Elizabeth Tupaj, Team 811 Crew.

That said, Another significant hurdle that AI could not solve was working with vendors. Students reported that while AI searches could find vendors, it couldn’t establish the necessary rapport or navigate tight timelines. Personal relationships proved crucial for securing parts and services.

The Indispensable Human Element: Judgment and Experience

Of the three finalist teams, only “Fast and Fractured” achieved first-attempt ignition of their mini-combustor, despite none of them having prior gas turbine experience. They used AI heavily for trade studies and architecture comparisons, arriving at a viable design.

Interestingly, The ultimate victors were Team 811 Crew, a more senior group with greater exposure to turbomachinery concepts. Their engine successfully started, transitioned to Jet-A, and generated net thrust. Interestingly, this team had been more resistant to using AI throughout the competition, relying instead on their foundational knowledge and teamwork.

“JARVIS taught me that getting value from AI takes two things: enough expertise to judge what it tells you and catch it when it’s wrong, and enough curiosity to actually lean on it where it could help.” – Professor Andreea Bobu.

The challenge highlighted a clear correlation: younger students, often with less foundational engineering experience, tended to use AI more frequently and creatively. More senior students, possessing deeper expertise, leveraged AI more judiciously or, in some cases, were more skeptical.

Lessons for the AI-Native Engineer

However, The JARVIS Challenge underscored that while AI can multiply engineering productivity, human judgment and first-principles thinking remain paramount. Mastering fundamental concepts and developing strong engineering judgment are critical for navigating complex decisions and incomplete information, especially when building safety-critical physical systems.

The implications for aerospace and other tough-tech industries are profound. If small teams, empowered by well-managed AI copilots, can compress design-build-test cycles from years to weeks, it could revolutionize R&D timelines, workforce structures, and competitive dynamics.

“JARVIS highlighted the power of AI in the design of physical systems. But it also showed that the key to unlocking that power is education, through coursework, internships, and hands-on extracurriculars… Performance in JARVIS correlated strongly with year in school. My main takeaway is that in the AI era, education is more valuable than ever.” – Zachary Cordero, Associate Director of the MIT Gas Turbine Laboratory.

Meanwhile, The students of the JARVIS Challenge are at the vanguard of a new era, grappling with the real-world stakes of AI integration in engineering, not as a theoretical exercise, but with a jet engine on a test stand.

Expert Perspective

From an industry angle, the clearest signal around AI in engineering design is how it may influence engineering. The story reads less like a one-day spike and more like a marker of broader movement.

The next phase will depend on how quickly teams, regulators, or customers react. In practice, that gives AI in engineering design room to reshape expectations across challenge over the near term.

For readers focused on practical impact, the best next step is to watch what changes around design once attention turns into execution.

Frequently Asked Questions

Why does AI in engineering design matter right now?

The Dawn of AI in Tough-Tech EngineeringAt a glance, Artificial intelligence, particularly generative AI and large language models (LLMs), has dramatically reshaped software engineering, automating code generation and vulnerability detection.

What broader change could AI in engineering design signal?

But how does AI fare when the challenge involves conceiving, designing, and fabricating complex physical systems, such as a jet engine?

What should the market watch next around AI in engineering design?

This was the central question behind the JARVIS Challenge (Jet-engine AI Research and Validation Intensive Sprint) at MIT.Meanwhile, The JARVIS Challenge aimed to determine if AI could significantly compress the traditional design-build-test cycle for safety-critical hardware.

Source: https://news.mit.edu/2026/can-ai-build-jet-engine-jarvis-challenge-tests-ai-copilots-in-tough-tech-engineering-0714