Jake Welde has joined Cornell Engineering’s Sibley School of Mechanical and Aerospace Engineering as an assistant professor, bringing with him a fascination for both the abstract mathematics that make robotics possible and the physical machines that bring those ideas to life.
Welde grew up in Stony Brook, New York, where his early interests took shape around mechanical play. “I spent all my time building little machines out of Legos,” he said. His first real taste of robotics came in 10th grade, when he competed in the Science Olympiad’s robot-arm event. “Before that I thought I was interested in engineering but wasn’t sure what area. After that, I knew I wanted to do robotics, and that’s stayed the same ever since.”
That curiosity eventually took him to the University of Pennsylvania, home to the GRASP Lab—one of the world’s premier robotics research centers. “That was the main reason,” Welde said. “I knew if I wanted to get a robotics-oriented education as an undergrad, I should go to a place where a lot was happening at the graduate level in robotics—and that turned out to be true.”
As an undergraduate, Welde joined the lab of Vijay Kumar, working closely with graduate student Justin Thomas on planning and control for drones equipped with arms—machines that could pick up objects or open doors. What began as hands-on experimentation soon evolved into a deep theoretical interest. “I realized that although I enjoyed the application, I was really drawn to the underlying algorithms and control theory,” he said.
That shift in focus carried through his doctoral work at Penn, where Welde began exploring how to generalize control algorithms across different robot morphologies. “Maybe someone designs a control algorithm for a particular robot,” he explained, “but then you add an arm, and everything breaks. I want to understand what, fundamentally, made the original method work so we can apply that idea to a broader class of robots.”
Initially, Welde expected to enter industry after earning his Ph.D. “I wanted to deepen my skills and go work for a company building cool robots,” he said. But as he dug deeper into theory, he found himself drawn to academia. “I liked the really abstract stuff—and I also realized I loved teaching.”
Welde discovered that designing assignments and working directly with students offered a refreshing counterpoint to the gradual, setback-prone pace of ambitious research. “Teaching lets you rediscover the things you love about the basics of your field and see them through fresh eyes,” he said. “It’s a really nice counterpoint.”
At Cornell, Welde will teach Dynamics, a sophomore-level course he previously helped teach at Penn. His lab, which still in the process of being set up, will eventually reflect his dual interests. “I envision my lab as being half math and half hardware,” he said. “You’ll see a spaces where we fly robots, build new morphologies, and develop control algorithms that can handle them.”
His research centers on aerial robots capable of both dexterous interaction and dynamic flight. “Today’s drones are agile but simple—they can’t really interact with their surroundings beyond collecting data,” Welde said. “Some aerial robots can grasp and manipulate, but they move slowly and quasi-statically. We want to build robots that can do both—interact richly with their environment and respond dynamically like animals in nature.”
To achieve that, his group will tackle questions that lie between control theory, morphology design, and AI. “Some students will work on the analytical side—generalizing algorithms for new robot forms,” he said. “Others might ask what the robot’s physical shape should be to enable new capabilities. And some will look at how we can make AI tools like reinforcement learning more efficient, so we can solve harder problems without vast computational resources.”
Welde’s interest in Cornell was partly professional and partly personal. “The department culture just seemed really healthy,” he said. “People are leading experts in their fields, but they’re also human beings. There’s a good balance here.” He also appreciated the school’s collaborative environment and the ease of working across disciplines. “My work sits at the intersection of many areas, and Cornell makes it easy to supervise students across different programs.”
Welde’s research group at Cornell is the Geometry, Design, and Control (“GeoDesiC”) Lab. He chose the acronym ‘geodesic’ because “it can be thought of as a ‘straight line’ through a ‘curved’ space, a concept closely related to the mathematical underpinnings of my work,” he said.
Outside the lab, Welde spends his time outdoors or in the kitchen. “I love to cook,” he said, “and I love being outside—spending time in nature is a big rejuvenator for me.” He also enjoys playing fetch with his dog every evening—an activity well-suited to his new home in Ithaca.
As he settles in, Welde looks forward to building a lab that welcomes students from across engineering and applied math. “I want to bring people with different skills together and see what we can create,” he said. “That’s where the most interesting ideas come from.”