Alumni push industry forward with robotic innovations
Robotics and artificial intelligence (AI) are two rapidly evolving fields, enabling machines to perform everyday tasks with incredible speed and accuracy. Now, two enterprising Buckeye engineers are leveraging this technology to grow their promising robotics companies and disrupt the industrial status quo.
Mechanical engineering alums Simon Kalouche ’14 and Andy Lonsberry ’13 recently returned to The Ohio State University College of Engineering to share the trials and tribulations of starting a new business, along with some tips for success. The two entrepreneurs presented at April’s Dave and Margie Williams Distinguished Lecture. The lecture series was established in 2019 by a generous gift from College of Engineering Dean Emeritus David B. Williams and his wife.
The future of fulfillment
Kalouche is the founder and CEO of Nimble Robotics, which engineers intelligent next-generation AI robotic fulfillment systems. Headquartered in San Francisco, Nimble’s vision is to build fully autonomous logistics so every brand can offer free delivery in two days or less.
“If you look at every fulfillment operation in the world today, from Instacart's grocery picking all the way to the most advanced Amazon warehouses, there are still millions of people required to manually pick and pack and handle your online orders,” said Kalouche. “No one has been able to fully automate the picking and packing step at scale.”
Because fulfillment services still have to depend on people for the bulk of the work, warehouse designs are limited to what humans can physically do, where they can go and what’s ergonomically safe. With robots that can pick, pack and handle millions of objects reliably, warehouses can be reimagined to unlock a simpler, smaller, faster and more efficient fulfillment center.
“That's what we're building at Nimble—a network of these next-generation, end-to-end robotic fulfillment centers where each one of these warehouses is performing fulfillment for dozens of brands at a cost and speed that traditional 3PLs can’t match.”
Kalouche’s love for robotics began his first year at Ohio State in the Fundamentals of Engineering for Honors program during which he had to design a robot that could navigate a specific course. His passion grew while completing an internship at NASA’s Jet Propulsion Lab (JPL), where he worked on a gecko adhesive climbing robot, and later while completing his master’s at Carnegie Mellon University, where he worked on dynamic jumping legged robots. But it wasn’t until Kalouche enrolled at Stanford University to pursue a PhD that he realized how AI could advance his robotics interests.
He began working on an algorithm called deep imitation learning, in which a person teleoperates a robot to perform a task it can’t yet do independently. The teleoperator remote controls the robot repeatedly on various tasks, and the recordings of those teleoperator actions become a data set which can then be used to train neural networks, the current state-of-art in AI. When trained on this teleoperator data, the robot begins to infer how it should handle different objects even if it hasn’t seen that exact item before.
“That's ultimately what led to Nimble being founded, as I saw potential to deploy that concept, that framework of teleoperating robots to provide real robotic labor today, collecting training data in the process and then using that data to train the robots to become fully autonomous in these traditionally un-automatable tasks over time to solve real world problems,” said Kalouche.
Kalouche put his PhD aspirations on hold in 2017 to start Nimble, which has been growing quickly over the last several years. The company began deploying its robots in 2019 at warehouses for national brands like Victoria’s Secret, Best Buy and Puma. Today, the company has three of its own fulfillment centers, with locations in the Bay Area, Texas and the East Coast, allowing them to reach 96% of the U.S. population in two days or less using ground shipping. And its robots? They’re more than 99% accurate, said Kalouche, and now rarely need help from humans. Over the next few years Nimble plans to expand its robotic fulfillment network with additionally nodes around the country and potentially Columbus one day as well.
As for the next crop of aspiring engineers turned entrepreneurs, Kalouche’s advice is to find a side project, an applied research project, a hands-on internship at a startup or a student organization, and “just get building.”
“My internship at JPL changed my world entirely. I was connecting the theory and what's in the books to how it's applied in the real world. I think that is so crucial and it will help everyone get more excited about learning and building.”
The future of manufacturing
Andy Lonsberry is co-founder and CEO of Path Robotics, which creates manufacturing robots that autonomously scan, position and weld parts without the need for skilled welders or robot programmers.
“We focus on manufacturing, and specifically the first vertical manufacturing that we deployed into is welding. The reason being is there's a massive labor shortage in the United States and globally, and it's continuing to grow,” said Lonsberry. “Government statistics show that there will be a labor gap of 400,000 human welders in the U.S. by 2024. It’s roughly a $26 million market opportunity.”
The Path system scans and creates a 3D model of each part. Its patented sensors are built to see highly reflective surfaces and survive harsh manufacturing environments. Using proprietary AI, the system analyzes the sensor data on the fly to understand each part individually. It adapts throughout the process, creating optimal robotic paths and part positioning to produce high-quality welds.
“Going forward, this is the full pipeline for a system that can see, do, learn and improve,” said Lonsberry. “We believe this is not just something that can be deployed for welding, but other tasks such as grinding, painting, assembly and so forth in the manufacturing industry.”
Lonsberry’s interest in robotics started in undergrad while he was interning at IHMC, a bipedal walking robot company.
“I got ingrained in seeing these humanoids come to life, seeing them take on really hard challenges, seeing them do things that I'd never seen another robot do,” he said.
One particular moment stands out in his memory—when a paraplegic patient was able to walk around the room while wearing the exoskeleton Lonsberry had been working on.
“It was incredible to watch. His family broke down. He broke down. It was such an amazing moment and really solidified in my mind that technology is here to help us as a human race continue to propel forward,” he said. “In that work is where I saw a lot of machine learning, a lot of AI, and recognized that I want to be a part of that journey.”
After graduating from Ohio State, Lonsberry earned his master’s in mechanical engineering from Case Western Reserve University and started his doctoral studies, but eventually dropped out to start Path with his brother in 2016. Based in Columbus, Path now employs 200 people and operates a 200,000-square-foot facility, which is helpful when welding large-scale items like a 100-foot utility pole electrical system or a 40-foot truck chassis. But according to Lonsberry, welding is just the beginning at Path Robotics as the company prepares to launch a new assembly product toward the end of 2023.
And while he doesn’t believe a graduate degree is necessary in order to find entrepreneurial success, he does think it can help students figure out what they’re passionate about.
“I wasn't successful in grad school relative to getting my degree, but I was able to utilize four years to get really good at a niche in the market at that time, which was deep reinforcement learning, and then utilize that to jumpstart my company,” said Lonsberry. “It was a great opportunity for me to explore a bunch of areas to also understand what I cared about most and what did I want to lean into further.”
by Meggie Biss, College of Engineering Communications | email@example.com