Learning About Robotics Helped These Humans Work on Autonomous Vehicles
When Ishan Khatri was little, he would pore over copies of Popular Science from the library, enamored with the futuristic vehicles, robots, and tools that filled the glossy, colorful pages, and the inspiring scientists, engineers, and designers making them.
Today, he’s one of those engineers at Argo AI, the self-driving technology company headquartered in Pittsburgh, Pennsylvania. He’s working on the technology that powers a kind or robot that the company is working to make widely accessible – the autonomous car.
And he’s far from the only one whose interest in robotics has helped them succeed in their work on autonomous vehicles. Argo AI’s CEO Bryan Salesky led software development for a team of Carnegie Mellon University’s (CMU) researchers that built a robot that won the Urban Challenge competition hosted by the U.S. Defense Advanced Research Projects Agency (DARPA). And prior to Argo AI, co-founder and president Peter Rander, worked at CMU’s internationally renowned National Robotics Engineering Center (NREC) and on commercializing robotics technology developed at the NREC.
We asked several Argo team members about how they got interested in robotics and how it helps them in their jobs advancing self-driving technology. Here they are in their own words, edited for length and clarity:
Rituja Marathe, Hardware Engineer, Lidar, Argo AI
“I first became interested in robotics when I grew up in India. My dad is a mechanical engineer who designs and manufactures tools for automation, which certainly led me to being interested in the field. I did an internship at a company called Bobst that does a whole bunch of robotics and automation.
During grad school, I also had the opportunity to join my school’s NASA Robo-ops team of 12 students. Essentially, teams work together to build their own prototype of the Mars Rover to go collect rock samples. It drives around in a Mars-like terrain in Houston, but is controlled remotely from across the country. I was primarily focused on the robotic arm that is supposed to maneuver and pick up these samples. It was really cool because I could do a lot of hands-on work. And it definitely piqued my interest in wanting to do robotics for my career.
At Argo, I work on the team responsible for working on the sensors that allow our autonomous vehicles to detect what’s around them and where they are moving. In this role, I work with a combination of optical, thermal and mechanical engineers. I also think about the processes that are needed to mass manufacture our sensors. When you’re making hundreds or thousands of multiples of a piece of equipment like a lidar sensor, it can be very helpful to use automation in that process.
For me, Argo is definitely a robotics company because the goal of robotics is to make a machine that makes human life better or easier. And with Argo’s focus on safety, we’re designing machines to make transportation safer, simpler, more efficient. We’re also a transportation company, but we’re not really building any vehicles, we’re building the brains of the vehicles.”
Ishan Khatri, Software Engineer, Special Projects, Argo AI
“When I was younger, my dad would always get Popular Science magazines from the library for me to read. There was always cool stuff in there and lots of it was robotics-related, but I distinctly remember one year reading about the DARPA Grand Challenge in the desert for self-driving vehicles.
There was a really large image of Carnegie Mellon University’s bright red Humvee somewhere in that article that’s totally ingrained in my memory. Ever since that point, I decided that I wanted to work on robots. I proceeded to do the FIRST Robotics challenge in middle and high school, major in computer science and eventually intern at Argo when I was a junior in college. The internship led to a full-time offer and now I’m quite literally working at my dream job.”
John Seminatore, Staff Technical Program Manager, Argo AI
“My previous experience was in aerospace engineering: I was an officer in the Air Force.
After the Air Force, I went to grad school at Virginia Tech University for mechanical engineering, and while there, I worked as a program manager for a team that participated in the DARPA Robotics Challenge building humanoid robots. We designed our robot from scratch, including everything – down to the actuators and joints.
It was a huge exposure to the multidisciplinary approach that robotics requires — the electronics, the software, the hardware — it’s a small-scale yet highly complex process. Robotics systems are very interconnected; if one system isn’t working well, you feel it, and then the whole system isn’t working well.
Most robots end up with the same breakdown of tasks that they need to do: See the world around them, understand it, decide what to do about it, plan how they move and interact. Some people tend to think of the cars we’re working on as cars, but we aren’t really building cars, we’re building robots.
Right now, the two main projects I’m focused on at Argo AI are highway testing and data logging. Until now, we’ve been very focused on driving in the urban core, so the highway testing is about figuring out how to handle higher speeds and lane changes safely. For the data logging, the challenge is making sure we collect and filter out the data we actually need in a format that we can understand. As we move to the next generation of sensors, we’ll be collecting even more data, so how do we get that stored efficiently?”
Neil Chauhan, Strategy Analyst II, Argo AI
“I went to Western University in Ontario and pursued mechatronic systems engineering. (It’s a multidisciplinary subject combining mechanical, electrical and software engineering.)
I began my professional career in the automotive sector, because that industry was where much of the mechatronic engineering was occurring professionally at the time. However, I initially found it to be more slow-moving than I would have liked: the industry is large and mature, and it typically takes between 5 and 10 years to create and release a new product.
Autonomous vehicles still seemed very far-off to most people. I had discussions with high-level decision makers in the industry, and they said that electric cars weren’t going to be possible until 2035 and autonomous vehicles not until 2070.
But in the mid 2010s, things changed with the advent of new products and startups. Suddenly, I was having very different talks with automakers who were more convinced that the technology for autonomous vehicles and electric vehicles was commercially viable and attainable sooner than they had first thought. That led me to seek out an opportunity at an autonomous vehicle tech startup, Argo.
My specific role at Argo is quite unique. I’m an engineering hire, but I work on the strategy team. My role involves communicating what’s special about Argo’s cutting edge technology to a range of audiences like investors, customers, partners, and our own engineering teams. I have a technical background that can help communicate to these audiences how the technical wins and progress we are making at Argo AI has a business value and economic value.
I now also spend considerable time thinking about how Argo can be positioned to provide the most value out to the world. Developing autonomous vehicles is considered one of the hardest robotics problems to solve, but one with tremendous economic value and business opportunity. As a result, there is funding and a drive to solve it.
Argo is listed as a transportation company on LinkedIn, but “mobility as a service” robotics is probably a better way to put it: we’re developing new ways to move people, goods, other things we can’t even imagine yet around in a safer and better way. Saying we are a transportation company probably underplays how strong technically we are. It’s not like we can’t take the problem-solving talent and apply it somewhere else.”