�ǿ����޴�ý

April 15, 2025

A team of four Olin College students and one Stanford undergraduate won the first-place prize at this year's The Generator AI , for Terraflex, their AI-powered prosthetic foot attachment. The students also took home the Nature-Inspired Design Prize (with a potential value of up to $60,000 for the team).

On Saturday, April 5th, Babson College's interdisciplinary AI Lab, , co-hosted an Entrepreneurship & AI Buildathon with the Cambridge, MA arm of Microsoft Research in the .

Over 200 students, forming more than 40 teams, from twelve universities (e.g. Babson, MIT, Northeastern, Olin, Harvard, Wellesley,) attended the one-day, intensive event, creating and pitching breakthrough new-venture concepts and prototypes for new AI-enabled products and services.

The Terraflex team, made up of Maya Adelman '27, Darian Jimenez '27, Jeffrey Woodyard '27, and Gabe Zak '25 from Olin, and Rhea Rastogi '28 from Stanford University, designed, prototyped, and presented Terraflex, an AI-driven prosthetic foot attachment that dynamically adapts to uneven terrain.

The varied technical skills of each team member, and the overall inspiration and collaboration of the group played a big role in the win.

"We got lucky with our team dynamic," says Maya Adelman, a sophomore mechanical engineering major. "Our work flowed seamlessly together, and each member brought completely different and essential skills to the table."

The students chose to compete in the Buildathon's AI-Powered Biomedical Device project track (3 different tracks were offered), and with multiple members of the team passionate about and experienced with assistive technology, it was a perfect fit.

Darian and Maya previously collaborated on a project creating a biomimetic prosthetic hand that mimics human movement through EMG sensors, and Maya has experience redesigning advanced wheelchairs in Tanzania, designing modifications to 3D-printed prosthetic hands for people around the world, and in designing and creating an ergonomic head/neck/shoulder support for a local wheelchair user in Massachusetts.

After establishing their collective interest in assistive tech, the Buildathon team got to work conducting research, and soon noticed a need that wasn’t being met with standard prosthetic leg designs: adaptability across natural, complex terrain.

"As we later highlighted in our pitch, over two million people in the U.S. live with lower-limb amputations, yet 60% cited an inability to walk on natural surfaces as a major limitation. We wanted to try to engage with an idea within this field that hadn’t been done before, which is how we arrived at Terraflex," says Rhea Rastogi, a freshman Biology + CS major at Stanford University.

Also, with this year's Buildathon focused on solving human design challenges, the Oliners were right at home.

"At Olin, we've had many opportunities to explore how engineers can draw inspiration from the world around them. Classes like Design Nature and have encouraged us to look into nature for solutions to some of today's most pressing problems," says Jeffrey Woodyard, a sophomore mechanical engineering major.

Next, the team looked at natural models that could inform the design of a prosthetic leg capable of adapting to uneven terrain.

That research led them to bipedal running birds—specifically the emu, ostrich, and roadrunner.

"We were especially intrigued by the roadrunner's zygodactyl feet (i.e. two toes facing forwards and two facing backwards, making an “X” shape). These feet are well known for helping the bird maintain balance and speed while running through rough desert terrain," says Jeffrey.

Inspired by the roadrunner, the five buildathoners began to develop a conceptual mechanical CAD prototype for a prosthetic foot attachment, utilizing a tension system for a multi-joint “talon” design that individually adapts to the terrain to optimize grip stability based on pressure sensor inputs. 

"We took the pressure sensors and placed them on three of Darian’s fingers to collect data on different balance states (forward leaning, backwards, stable, etc.)," says Gabe Zak, a senior mechanical engineering major.

"We then trained and saved a model on the data and then used that to make a prediction in real-time. We also had a switch to detect when the foot was on the ground so it wouldn't make predictions if the user was actually walking," he added.

Asked what they might add to the prototype in the future when they have more time, Darian Jimenez, a sophomore engineering major offered, "Many more sensors and a way to adjust the foot based on balance. We were thinking of adding other functionality such as muscle activation detection, but we unfortunately didn't have enough time to implement these features given the competition timetable."

The Buildathon timetable gave participants just a few hours to complete all the work needed to bring their AI-driven products, solutions, and services to life, and while an inward confidence may have been felt by the team after they had the concept and their strategy for creating the prototype (and much of the AI component of the project was familiar content they said, learned in "Machine Learning," a course taught by Olin professors Sam Michalka and Paul Ruvolo), outwardly, they were scrambling to race against the clock.

"We were only given around 5-hours to complete all the work, which was squeezed even further by the mechanical component of the aspect, as we had to CAD in record time to allow for ample 3D printing," says Maya.

Because they compressed so much technical work into the build, the team didn’t have time to rehearse their pitch. So, they just went for it.

"We never expected to win in the beginning—there were around 40+ teams with 12 prestigious schools involved, and two competitive first-place prizes. We ended up having several organizers of the event come up to us to praise our project and predict our win after the second round of pitches consisting of the top 8 finalists. The reality of winning became way more real after that," says Maya.

The was $2,500 per team and the bonus biomimicry prize - which the judges said the group won by thoughtfully incorporating inspiration from the zygodactyl feet of roadrunner birds into their design, is valued at a total of up to $60,000 for the team. This consists of an asynchronous biomimicry certification program valued at $5,000 per recipient and a 4-8 week summer internship opportunity valued at $2,700 per week per recipient.

The main monetary fund that will take the Terraflex prototype and idea to the next level, however, is an offer the students received from one of the judges to help fund the project for future work.

"We are considering spending our next semesters at Olin developing the project further and making a more realized device, although nothing is confirmed yet at this stage," says Maya. 

Additionally, the team recently applied to and got accepted to the  set to take place on April 29, 2025, from 4:30 - 6:30 P.M. in the Knight Auditorium. The event is promoted as a chance to see cutting-edge ideas in action, and to gain new insights.