Charlie Kemp and Aaron Edsinger won the best paper award at the 2006 IEEE-RAS International Conference on Humanoid Robotics for a system that let the humanoid robot Domo help a person put objects on a shelf. That result pushed Kemp toward assistive robots, and it led to a faculty position at Georgia Tech. Today he is cofounder and chief technology officer of Hello Robot, which builds non-humanoid mobile manipulators for homes and workplaces.
From humanoid research to Hello Robot
Kemp became interested in artificial intelligence as an undergraduate at MIT. Rod Brooks had just started the Cog project, which used humanoid robots to investigate embodied intelligence. Kemp later joined Brooks's lab and worked with Edsinger on Domo.
The shelf-stocking paper led him to assistive robotics. At Georgia Tech he ran the Healthcare Robotics Lab, which focused on physical assistance with feeding, hygiene and dressing. He and Edsinger now lead Hello Robot, which they founded together. Kemp says he would not have predicted that chain of events.
His current focus is mobile robots with arms that benefit people in homes and workplaces. Many of the applications he points to involve close physical contact between a person and a machine, such as robots that assist older adults and people with disabilities. He ties the work to a practical concern: finding human assistance is becoming harder, in part because societies are aging. In his view, robots could improve quality of life and give people more independence.
Learning from the people who use the robots
Kemp says his best working days center on the people using a robot, not the robot itself. He names Keith, Henry Evans, Jane Evans and hundreds of others who have worked with robots from Hello Robot and his academic lab as having changed how he thinks about robotics.
Those users have shaped his design questions. Sometimes what matters most to a person is scratching an itch or picking up a dropped object. Sometimes autonomy is not the answer, because the individual wants more control. Kemp says answers like these do not come from programming alone in a room. Real-world use almost always reveals new opportunities for improvement, and he says those lessons shape the robots his team builds.
Hello Robot also collaborates on assistive robots for older adults and for people with severe mobility impairments. Kemp says he values the community that uses Stretch robots as research and development platforms, and he says those users keep finding new ways human-friendly robots can benefit society.
Robots for Humanity and Stretch
Kemp singles out Robots for Humanity as his favorite project. In 2010, Henry Evans saw a demonstration of the Georgia Tech lab's research, and that began a collaboration that continues today. Henry Evans has severe quadriplegia. Kemp describes him and Jane Evans as pioneers in assistive robotics.
Working with Henry, Jane and others, the Healthcare Robotics Lab made progress toward mobile manipulators that can meaningfully benefit people with disabilities in daily life. Building a home robot for the Evanses was the main inspiration for the prototype that led Kemp and Edsinger to found Hello Robot. Henry and Jane have used each generation of Stretch, and Kemp says their experience influenced later designs, especially Stretch 4.
One episode surprised him. Henry Evans used Stretch to present a rose to Jane. The motions involved were relatively simple, Kemp says, but the gesture was emotionally powerful. He had expected his work on feeding, hygiene and dressing to be about physical help. He now sees that physical assistance can also help people connect emotionally.
Advice for newcomers
Kemp warns that online robot videos can make robotics look solved. He says a demo video differs greatly from a robot being productive in the real world, and that the world is very good at breaking robotic systems in ways that surprise roboticists. He still sees an enormous amount left to build, learn and discover, and he says newcomers should not feel behind.
He also says a robotics career does not require a focus on robotics technology. The most impactful future applications, he expects, will depend on technologists working with experts in specific application areas. His example is Dr. Vy Nguyễn, Hello Robot's clinical research lead, who began in robotics while earning a doctorate in occupational therapy and whose expertise has shaped the company's assistive work.
Analysis: what this does and does not show
The clearest lesson in Kemp's account is about method. The design questions he lists, such as whether a person wants an itch scratched, an object retrieved or simply more control, came from long work with specific users, not from a technical roadmap. Henry and Jane Evans have used every generation of Stretch, and Kemp credits that use with shaping the later designs. That is a direct feedback loop between disabled users and a commercial product line, and it is a stronger basis for assistive robots than lab testing alone.
It also reflects a broader split in the field. Kemp's career began with humanoid research at MIT, yet his company builds non-humanoid mobile manipulators. That is a choice about form that follows from the tasks he describes, which are fetching, handing over and assisting at close range. The source does not compare Stretch with other platforms, so it does not show that this form is better than alternatives. It shows only that this team moved away from humanoids for this work.
Much remains unproven. The source gives no specifications, prices, user counts or trial results for Stretch, and it offers no measurements of independence or quality of life. The claim that robots can empower people to live more independently is a stated expectation, not a demonstrated outcome. Kemp's own caution points the same way: he says the gap between a demo and productive real-world use is vast. The rose handed to Jane Evans is a meaningful anecdote, not evidence that such robots work reliably in unsupervised homes.
The open questions are the ones Kemp raises himself. How much control users want, and when autonomy helps or gets in the way, will likely vary person to person. Whether assistive robots can meet the needs of an aging population depends on results the source does not report. What it does establish is who is shaping the work, and that clinical expertise sits inside the company alongside the engineering.