MIT’s JARVIS Challenge: AI Accelerates Jet Engine Design, But Engineering Judgment Remains Paramount
Artificial intelligence has demonstrably revolutionized software engineering, enabling the rapid generation of code and documentation, and enhancing performance monitoring and security analysis through sophisticated machine-learning algorithms. However, the application of these powerful AI tools to the intricate and safety-critical domain of conceiving, designing, and fabricating complex physical systems, such as a jet engine, presents a more nuanced challenge. This past semester, the Jet-engine AI Research and Validation Intensive Sprint (JARVIS Challenge) at MIT sought to explore this frontier, tasking undergraduate students with determining if AI can significantly compress the traditional design-build-test cycle, ultimately empowering them to build faster and more effectively.
The overarching conclusion from the JARVIS challenge, according to Professor Zolti Spakovszky, director of the MIT Gas Turbine Laboratory, is that AI can indeed substantially accelerate safety-critical hardware engineering. Yet, he emphasizes, "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. Manufacturing – not engineering design or analysis – remained the fundamental rate-limiting step." This sentiment underscores a critical finding: while AI can serve as a powerful co-pilot, the ultimate responsibility and strategic direction lie with human expertise.
The Challenge: A Race Against Time and Technology
The JARVIS Challenge presented a demanding four-week timeframe for undergraduate teams to design, fabricate, assemble, and rigorously test a small gas turbine aero engine. The core objective was to construct a "JARVIS-class" single-spool jet engine capable of producing between 50 and 100 pounds of thrust, operating on Jet-A fuel, and successfully completing five 60-second test runs. Teams were granted complete autonomy in their design choices, material selections, and fabrication methods, fostering an environment of innovation and rapid iteration.
Thirty-one students, representing a diverse spectrum of engineering disciplines across MIT’s School of Engineering, formed seven distinct teams. The composition of these teams varied significantly, from groups of first-year students with limited prior exposure to advanced engineering concepts, to more senior-heavy cohorts. A notable aspect of the challenge was the diverse experience levels; many competitors initially possessed little to no background in turbomachinery, compressible flows, or even fundamental thermodynamics, particularly among the younger participants. For many, this was their first encounter with the internal workings of a gas turbine.
Tools of the Trade: A Blend of Traditional and Cutting-Edge Resources
Participants had access to MIT’s state-of-the-art machine shops and a network of manufacturing vendors. Complementing these physical resources were industry-standard commercial software packages, including Concepts NREC for turbomachinery design, SolidWorks for 3D modeling, and ABAQUS for finite element analysis. Various specialized test rigs were also available for characterizing and assembling individual engine components.
Crucially, the teams were provided with access to MIT Parley, a newly launched platform designed to aggregate frontier large language models (LLMs) through a unified interface. This innovative tool allowed JARVIS leadership to monitor student engagement with AI in real-time, tracking prompts, associated costs, specific LLMs utilized, and other vital data points. With early access to Parley secured for all participants and substantial financial backing from MIT Lincoln Laboratory, the Department of Mechanical Engineering, and key corporate sponsors—Safran, Voyager Technologies, Beehive Industries, and Boom Technology—students enjoyed virtually unlimited access to AI resources.
The sponsorship of the JARVIS Challenge was driven by a dual interest: a keen eye on future recruitment and a genuine curiosity about the transformative potential of AI in reshaping engineering workflows. Ryan (Hal) Hefron of Voyager Technologies articulated this sentiment, stating, "We see this as the future of engineering. You’re honing skills that are not just nice to have – they’re going to be the future baseline in the engineering workforce."
Vincent Garnier, managing director of Safran Tech, observed the competition with keen interest, characterizing JARVIS as "a genuine experiment, a learning endeavor." He expressed his initial uncertainty about the outcomes from both the students and the AI models. "What struck me coming from the students was: first, the enthusiasm to explore; then, as the project developed, they all came to the cool-headed realization of what AI could or could not help them with, and then almost instantly adapted for that," Garnier commented. He further noted, "It makes me confident that this generation of leading engineers will probably not fall prey to easy and shortsighted use of AI, and will do so by keeping ever more in contact with experiments – physical or thought experiments."
Faculty leadership, including Professors Zachary Cordero, Zolti Spakovszky, Masha Folk, and Andreea Bobu from the Department of Aeronautics and Astronautics, alongside engineers from Lincoln Laboratory and a dedicated team of teaching assistants, provided essential oversight to ensure safety and guide the learning process. Regular weekly progress reviews involved critical evaluation of student advancements and their strategic utilization of AI tools. Professor Spakovszky developed a nuanced approach to mentorship, aiming to guide teams without providing direct solutions. His feedback often involved posing incisive questions, such as, "Do you know what a rabbet fit is? Take in the comment," prompting students to delve deeper into their understanding and research.
The AI Divide: Where Assistance Flourished and Limitations Emerged
By the end of the first week, one team had withdrawn from the competition, while the remaining teams had, with varying degrees of success, developed initial design concepts for their gas turbines. AI proved invaluable in numerous capacities: summarizing technical literature, providing tutorials for design software, identifying potential vendors, generating comparative analyses for design decisions, and answering specific technical queries. One team even deployed an AI agent within Parley to function as their project manager, demonstrating a creative integration of AI into team operations.
The second week marked a pivotal shift as teams transitioned to detailed CAD design, part procurement, and the prototyping of their combustors. It was at this juncture that the inherent limitations of AI began to surface more prominently. While LLMs like Claude and ChatGPT excelled at proposing design alternatives and bridging knowledge gaps, the well-documented issues of AI "hallucinations," sycophantic responses, and a fundamental lack of intuitive physical understanding started to erode student confidence and impede progress.
Elizabeth Tupaj, a member of the 811 Crew team, articulated this challenge: "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."
Teaching assistant John Zhang observed a recurring pattern: "Seeing this firsthand with the students reminded me how much first impressions matter. If the students couldn’t get answers from the AI early on, they quickly grew frustrated and formed a lasting opinion that precluded them from using it later." This highlights the critical importance of early, successful AI integration for sustained adoption.
In the critical final weeks, the leading teams encountered a formidable obstacle that no AI could overcome: navigating the complexities of vendor relationships. Students reported, "AI searches found vendors we had no rapport with, who had no interest in our tight timeline. The vendors who came through were the ones our team had personal relationships with." This underscores the enduring value of human networking and established professional connections in practical engineering endeavors.
Among the three finalists, only the "Fast and Fractured" team achieved first-attempt ignition of their mini-combustor. This team had extensively utilized AI for trade studies and architectural comparisons, successfully arriving at a viable design despite no prior gas turbine experience among its members.
Masha Folk, the Charles Stark Draper Career Development Professor of Aeronautics and Astronautics, reflected on the project’s success: "The JARVIS Challenge showed what’s possible when you combine AI-enabled design with motivated students and a culture of rapid experimentation. The moment that stood out most was when the first student-designed combustor was installed on the test stand. It ignited flawlessly, ramped to full power, transitioned to dual-fuel operation, and then sustained stable combustion on 100 percent Jet-A fuel. This was proof that we can dramatically accelerate the cycle of design, build, and test while giving students hands-on experience with a real engineering challenge."
At the Vanguard of AI-Native Engineering
By the end of May, two more senior teams, "Fast and Fractured" and "811 Crew," had successfully completed full engine tests. "Fast and Fractured," despite their AI-assisted design prowess, faced significant delays due to vendor issues but ultimately reached the testing phase. Their engine test, however, was prematurely halted when the rotor experienced contact and seized against the stationary housing. In contrast, the "811 Crew," who entered the competition with a stronger foundation in turbomachinery and propulsion concepts, emerged as the victors. Their engine successfully started, transitioned to Jet-A fuel, and generated net thrust, marking a significant accomplishment.
PhD student Joe Chiapperi described the palpable tension during the tests: "As we stood there with the air-starter, hearing their engines spool up and watching them spit fire, it felt like my heart was racing out of my chest. There were so many ways it could go wrong! What these students accomplished in such a short time span is nothing short of amazing."
Intriguingly, the winning "811 Crew" team had initially been hesitant to heavily rely on AI, prioritizing their foundational knowledge and teamwork. Tupaj explained, "We had people who were at least somewhat familiar with the design software, mechanical engineers who knew how to build anything, and aerospace engineers who had taken classes on the design of gas turbine engines specifically." This strategic choice, rooted in existing expertise, proved to be a winning formula.
Analysis of team performance revealed a correlation between AI utilization and experience. Younger students tended to leverage Parley more frequently and with innovative approaches, while juniors and seniors drew upon their deeper domain knowledge. Professor Andreea Bobu offered a profound insight into this dynamic: "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. The team that moved fastest in the sprint was experienced and leaned heavily on AI to get there. The team that eventually won was more resistant to AI; they had the expertise, but that skepticism made them slower. The sweet spot seems to be knowing enough to stay in charge of the tool, and being eager enough to pick it up in the first place. To me, that’s the real opportunity ahead: training the next generation of engineers who have the judgment to direct these AI tools and the instinct to reach for them."
The competition’s most salient takeaway is that engineering experience acts as a powerful multiplier, and the human element remains indispensable. A robust understanding of first principles and fundamental concepts cultivates sound engineering judgment, enabling engineers to navigate complex decisions amidst incomplete information. When it comes to constructing safety-critical physical systems, the article strongly asserts that human hands and human accountability are irreplaceable.
"JARVIS has shown that AI copilots can have a multiplicative effect on engineering productivity, with judgment and first-principles thinking serving as the key differentiators among teams," added teaching assistant Kyle Woody.
Broader Implications for Aerospace and Engineering Education
The implications of AI’s integration into aerospace engineering are profound. If small, agile teams leveraging well-managed AI copilots can dramatically compress design-build-test cycles from years to mere weeks, the ramifications for workforce structure, research and development timelines, and competitive dynamics within the industry could be substantial. The students who participated in the JARVIS Challenge are at the forefront of this transformation, grappling with these stakes not as theoretical exercises, but within the tangible reality of a machine shop with a jet engine on the test stand.
Professor Cordero, associate director of the MIT Gas Turbine Laboratory, concluded, "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 like MIT Motorsports and Rocket Team. Performance in JARVIS correlated strongly with year in school. My main takeaway is that in the AI era, education is more valuable than ever." This sentiment reinforces the critical role of educational institutions in preparing the next generation of engineers to effectively harness AI while maintaining a strong foundation in core engineering principles. The JARVIS Challenge serves as a compelling case study, illustrating that the future of complex physical system design lies in a synergistic partnership between advanced AI capabilities and the enduring wisdom of human engineering expertise.
