The JARVIS Challenge, an intensive sprint at MIT, sought to answer a fundamental question: can artificial intelligence, which has revolutionized software engineering, achieve similar breakthroughs in the design and creation of complex physical systems like jet engines? This past semester, the initiative tasked MIT undergraduates with a demanding objective: to conceive, design, fabricate, assemble, and test a small gas turbine aero engine within a compressed four-week timeframe, leveraging AI as their primary engineering partner. The results offered a nuanced perspective, demonstrating AI’s capacity to accelerate safety-critical hardware engineering while underscoring the indispensable role of human engineering judgment.
The Genesis of the JARVIS Challenge
The JARVIS Challenge (Jet-engine AI Research and Validation Intensive Sprint) was conceived by the MIT Gas Turbine Laboratory, under the directorship of Professor Zolti Spakovszky, as a novel pedagogical experiment. The goal was to push the boundaries of traditional engineering education by integrating cutting-edge AI tools into the physical design-build-test cycle, a process that typically spans years. The initiative aimed to not only assess the efficacy of AI in accelerating complex engineering tasks but also to cultivate a new generation of "AI-native" engineers, defined not by their mere use of AI, but by their ability to lead and critically evaluate its outputs.
The context for the challenge is the rapid advancement of generative AI and large language models (LLMs), which have demonstrated remarkable capabilities in generating vast quantities of code and documentation. Machine learning algorithms are also proving invaluable in monitoring system performance and identifying security vulnerabilities. However, the leap from the digital realm of software to the tangible, multifaceted demands of physical engineering – particularly in safety-critical applications like aerospace – presented a unique set of challenges and opportunities. The JARVIS Challenge was designed to bridge this gap, offering a microcosm of the potential and limitations of AI in this domain.
The Teams, the Tools, and the Ambitious Task
The core objective of the JARVIS Challenge was for student teams to design and build a "JARVIS-class" single-spool jet engine. This miniature marvel was required to produce between 50 and 100 pounds of thrust, operate on Jet-A fuel, and successfully complete five 60-second test runs. The students were granted complete autonomy over their design choices, material selection, and fabrication methods, fostering an environment of innovation and independent problem-solving.
Thirty-one students, representing nearly every department within MIT’s School of Engineering, formed seven distinct teams. The composition of these teams varied significantly, ranging from groups of first-year undergraduates to more senior-heavy collaborations. A common thread among many participants was a limited prior experience in specialized fields such as turbomachinery, compressible flows, or even fundamental thermodynamics, particularly for the younger students. Many had never before encountered the inner workings of a gas turbine engine before committing to building one. This diverse background presented both a steep learning curve and an opportunity for rapid skill acquisition.
The students were provided with access to MIT’s state-of-the-art machine shops and a network of trusted manufacturing vendors. Commercially available software, including industry-standard tools like Concepts NREC for aerodynamic design, SolidWorks for CAD modeling, and ABAQUS for finite element analysis, were made available. Additionally, various test rigs were on hand for the precise characterization and assembly of individual engine components, ensuring a controlled and rigorous testing environment.
A pivotal element of the challenge was the integration of MIT Parley, a newly launched platform designed to aggregate a suite of frontier LLMs through a unified interface. This provided the JARVIS leadership with unprecedented visibility into the students’ AI utilization, allowing them to monitor prompts, associated costs, the specific LLMs employed, and other critical data points. The JARVIS leadership secured early access to Parley for all participants. Bolstered by significant financial backing from MIT Lincoln Laboratory, the Department of Mechanical Engineering, and corporate sponsors including Safran, Voyager Technologies, Beehive Industries, and Boom Technology, the students enjoyed virtually unlimited access to AI resources.
The sponsorship was driven by a dual motivation: a keen interest in recruitment and a genuine curiosity about how AI might fundamentally reshape engineering workflows in the aerospace sector. Ryan (Hal) Hefron of Voyager Technologies articulated this sentiment to the students, 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. He described the JARVIS Challenge as "a genuine experiment, a learning endeavor. We frankly didn’t know what to expect, from the students or from the AI models." Garnier noted that the students’ initial enthusiasm for exploration evolved into a "cool-headed realization of what AI could or could not help them with, and then almost instantly adapted for that." He expressed confidence that this generation of engineers would avoid "easy and shortsighted use of AI" by maintaining a strong connection with experimental processes.
Faculty leadership, including Professors Zachary Cordero, Zolti Spakovszky, Masha Folk, and Andreea Bobu from the Department of Aeronautics and Astronautics, alongside engineers from MIT Lincoln Laboratory and a dedicated team of teaching assistants, ensured the safety and integrity of the project. Weekly progress reviews provided a critical forum for faculty to assess student advancements and their strategic application of AI tools. Professor Spakovszky developed a unique mentorship approach, guiding teams with insightful questions rather than direct answers, prompting critical thinking and self-discovery.
AI’s Dual Role: Aiding and Hindering Design Processes
By the end of the first week, one team had withdrawn, but the remaining groups had, with varying degrees of success, developed initial designs for their gas turbines. AI proved to be a versatile tool in these early stages. Students employed LLMs to summarize technical textbooks, learn new design software, identify potential vendors, generate comparative analyses for design decisions, and source relevant academic and technical references. One team even established an AI agent within Parley to function as their project manager, automating certain organizational tasks.
The second week marked a critical transition as teams began to focus on detailed CAD designs, sourcing components, and prototyping their combustors. It was during this phase that the limitations of AI in physical engineering became more apparent. While LLMs like Claude and ChatGPT offered valuable design alternatives and helped bridge knowledge gaps, the well-documented issues of AI "hallucinations" (generating plausible but incorrect information), sycophancy (agreeing with user input even when incorrect), and a fundamental lack of embodied, physical understanding began to undermine student confidence and slow progress.
Elizabeth Tupaj, a member of the team "811 Crew," reflected on this experience: "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 that these initial frustrations could significantly impact long-term AI adoption. "Seeing this firsthand with the students reminded me how much first impressions matter," he noted. "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."
The final weeks of the challenge presented another significant hurdle that no AI could surmount: navigating the complexities of vendor relationships and manufacturing. Students reported that AI-driven vendor searches often yielded contacts with whom they had no established rapport and who were unwilling or unable to meet the project’s tight deadlines. "AI searches found vendors we had no rapport with, who had no interest in our tight timeline," students recounted. "The vendors who came through were the ones our team had personal relationships with." This highlighted the enduring importance of human networking and negotiation in the physical product development cycle.
Of the three teams that reached the finals, only "Fast and Fractured" achieved first-attempt ignition of their mini-combustor. This team had extensively utilized AI for trade studies and architectural comparisons, enabling them to arrive at a viable design despite having no prior gas turbine experience.
Masha Folk, the Charles Stark Draper Career Development Professor of Aeronautics and Astronautics, emphasized the synergy of AI-enabled design and student motivation: "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."
The Vanguard of AI-Native Engineering: Lessons Learned
By the end of May, two teams, "Fast and Fractured" and "811 Crew," successfully completed full engine tests. "Fast and Fractured," despite their AI-assisted design prowess, faced significant delays due to vendor issues. Their engine’s hot fire test was ultimately cut short when the rotor experienced a rub and seized against the stationary housing.
The victorious team, "811 Crew," benefited from their members’ pre-existing familiarity with turbomachinery and propulsion concepts. Their engine successfully started, transitioned to Jet-A fuel, and generated net thrust, securing them the win. PhD student Joe Chiapperi described the tension of the final 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."
Interestingly, the "811 Crew" had initially been more resistant to leveraging AI throughout the competition, relying instead on their foundational knowledge and strong teamwork. "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," explained Tupaj.
A key observation from the challenge was the differential use of AI based on student experience. Younger students tended to utilize Parley more frequently and creatively, while juniors and seniors leaned more heavily on their existing expertise. Professor Andreea Bobu summarized 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."
Bobu further elaborated on the optimal approach: "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."
Broader Implications for the Future of Engineering
The most significant takeaway from the JARVIS Challenge is the enduring value of engineering experience as a multiplier. The human factor remains an indispensable element in complex design processes. Mastery of first principles and fundamental concepts not only fosters sound engineering judgment but also equips individuals to navigate challenging decisions amidst incomplete information. In the realm of safety-critical physical systems, human oversight and accountability are irreplaceable.
Teaching assistant Kyle Woody articulated this point: "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."
The implications of AI in the aerospace industry are profound. The possibility for small teams, empowered by well-managed AI copilots, to compress design-build-test cycles from years to mere weeks could trigger substantial shifts in workforce structures, research and development timelines, and competitive dynamics across the industry. The students who participated in the JARVIS Challenge are at the forefront of this transformation, experiencing the stakes not as a theoretical exercise but within the practical context of a machine shop and a jet engine undergoing rigorous testing.
Professor Zolti Spakovszky concluded that the challenge illuminated the power of AI in designing physical systems. However, he stressed 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."
The JARVIS Challenge stands as a testament to the evolving landscape of engineering education and practice, underscoring a future where human ingenuity, critical thinking, and the judicious application of artificial intelligence will converge to drive innovation in complex physical systems.
