Interview Transcript
Benny: It is a pleasure to have Paolo Pirjanian, CEO and Founder of Embodied, with us today to talk about compassion, technology, and leadership! We thank you Paolo for taking his time sharing his insights with us.
Part 1: Leading with Compassion
Paolo: Leading with compassion doesn’t mean doing what they want, but doing what they need. Sometimes, as a leader, you could be making the least popular decisions. You make a decision that you think it’s going to be beneficial to your team. For example, over the last 2 to 3 years, we have to make some tough decisions. Our team wanted to hire more people and spend more money. At this stage of our development, it is important to spend money for the biggest bang for the buck. One of my management principles: “spend money where you can reduce uncertainty.” When you have a better understanding of the path forward, then you can put more fuel to the fire. When the situation is not clear, you have to go slower. When you have clarity, then you can put more gas and go much faster. In the past, our team pushed me with lots of arguments why we need to hire more. “If we are not hiring, we can’t do this!” Many times I had to push back by saying, “Let’s get over the following hurdles. I need to have more confidence in our ability to reach our goals. If not, we will be out of money and not able to get to the finish line.” I like to preserve cash until we know how to burn it. In retrospect, the team appreciates that we increased our runway by more than a year. It wasn’t an easy decision but it was the right decision. That’s an example of doing the right thing for the team by taking their input along with your experience to decide what’s the right way to go.
“Doing the right things for the team - not necessary what they want.”
I could be the most popular CEO. But, if we fail, we fail. No one would be happy! Our employees, our shareholders, and our customers would not be happy. Leading with compassion is personally hard when your team doesn’t necessarily understand. It is very tough to resist at first. After enough discussions, at some point, I have to say, “listen, you got to trust me on this one.” It is important that they commit even though they might disagree with the decision.
Benny: How would you describe the decision process in your company?
Paolo: First, we guarantee that we would listen to all input with an open mind. Second, we would spend time understanding the different perspectives. Third, we do not guarantee that this would be a committed decision. At the end of the day, whoever is responsible, whether it is the manager of the group or the head of the company, he/she will make the ultimate decision with input from everyone. But, the decision may not be the one you are hoping for. We expect everyone to commit to the path forward despite disagreements.
Our decision making process: listen, discuss, decide, and commit
If we have a manager that continues to make bad decisions, that person is not going to stay as a manager too long. If we put someone in a leadership position in the company and they keep making bad decisions, the team is looking at you, the leader, to make a correction. Otherwise, you would lose credibility with the team as well.
Benny: Great point. This is also true for us as individuals. If we make bad decisions, we would make corrections ourselves.
Paolo: Owning up and explaining how you are going to make corrections is important. I have hired direct reports to me that turns out to be not the right hires. If I don’t make the corrections there, the team is going to lose confidence in me as the leader.
Benny: When you discovered you hired the wrong person, what actions did you take and how long did it take you to make the corrections?
Paolo: Always take too long! This is an achilles’ heel for me and many leaders. When we reflect on firing, we realize we should have done it earlier. My personal belief is this: when I hire people, I give them 100% of my trust and confidence to do their job along with any support they need to succeed. When I realize that they are not necessarily doing a good job, I will try to spend a lot of time guiding them and coaching them in resolving the challenge they have been facing. This is the part where I struggle because I usually spend a lot of time. I usually spend about a year from the time I realize there is a problem with an employee when they might not have the right skills. I would spend a year training and coaching them. In some cases, it works out. The person responds really well. Which is great. They become stronger leaders. And, in some cases, it doesn’t work out.
Part 2: Examples of compassion in management
Paolo: When we have an exit conversation, I would tell them that, we like you. We have been very supportive of you. We have been trying to help you to overcome the following things that you are struggling with. We have come to realize that we have not made enough progress. We are out of suggestions and tools. What are your suggestions? Is there anything you think we can do? If the person doesn’t have any suggestion, he/she usually would conclude that they would move on. We are supportive of people moving on if there is no better suggestion.
Benny: A year seems like a long time.
Paolo: It is. Humans are involved. Families are involved. We want to make sure we have done everything we can to make the person successful. It depends on the candidate as well.
Benny: Is there a possibility that they are in the wrong role?
Paolo: Yes. Part of the feedback is to focus them on something else. You rotate people to see if they can find the right spot. Sometimes, it works out. I had a strong individual engineer who also demonstrated strong leadership skills. We started elevating the person to a leadership role like the head of engineering. Then, the person started failing. Things were falling behind and things weren’t getting done. I had this conversation with him, “Look. You are amazing! Thank you for taking on the challenge of leading the engineering team. But, we feel that you are spread too thin and your core work is suffering.” We discussed whether it makes sense for him to go back to the area where he started. At first, he was disappointed. But, then he realized he was being put in a position where he would fail.
Benny: Some people might need the training to scale. How do you go about training people as the company scales?
Paolo: That’s also what we decided to do. We are not saying he cannot get this position again. We are going to spend additional time training and helping him in order to acquire the skills required to succeed in that role. He is now thriving in the company.
Benny: When you have different people with different experience and skill sets, it can help them to learn quickly.
Part 3: From smuggling across the border to starting up
Benny: How about your story? Please share with us how you came to this point.
Paolo: I was born in Iran as an Armenian which is a minority Christian nation. After the revolution in 1979, it was not safe to stay in the country as a religious minority. So, my dad decided in order to have any reasonable prospect in education, he wanted to get us out of Iran. Which at the time due the war between Iraq and Iran, it wasn’t possible to do it legally. So, we smuggled through the mountains out of Iran. We went through Turkey and became refugees in Denmark. That journey through the mountains took about 8 days walking. We stayed in Turkey for another month and a half. Then, we found another group of smugglers to take us to Denmark. The reason we had to get out of Turkey is because there was a bad history between Armenian and Turks. If the Turkey government found out about Armenians being in Turkey illegally, that wouldn’t have ended well either.
Startups require the ability to deal with uncertainties a lot. When I was 16 going through the mountains, ended up in Turkey and figured our way to Denmark, we did something we never did before. So, you have zero knowledge about the task, we had to figure it out on the spot. I think it sharpens your instinct. That’s survival! Potential death would be the outcome of that journey. We had to follow our instincts and reacted to very limited information without any knowledge of the tasks. So you got to learn quickly, you got to make decisions based on incomplete information, and I think that sharpens my instinct in making decisions in an uncertain environment with very limited information, which is very similar to what you need to do in a startup. In a startup, you have an idea of what you want to do. But, you have to figure out how to go about it.
Part 4: His journey to engineering
Paolo: In terms of technology, I have always had a strong aptitude towards technology since my early childhood. When we ended up in Denmark, I learned the local language. At first, I wanted to be a doctor because immigrant parents wanted kids to be doctors or lawyers. It was perceived as the secure path to a decent living. I didn’t know why other than my parents wanted me to be a doctor. During the 2nd year in high school, I saved some money for a week-long vacation in Spain during the summer. I bought the plane ticket but the travel agency informed me that I needed to get a visa because they just changed the law. They asked me to call the Spanish Embassy. But, I was told there is no way to issue it in one week. So, I couldn’t go. The travel agency was kind enough to reimburse the ticket. It was about $500. As a teenager, I was thinking I needed to do something useful with this money. So, I went and bought a computer without knowing what it was. At the time, the most popular computer was the Commodore 64. I went to the store and it was too expensive for me. Then, I saw this flat and black computer. It wasn’t round or beige. The price was slashed to $400 from $800. I asked if this computer was used. “No, it is brand new.” So, I bought it and took it home. I had no idea what to do with it. It is like an alien coming to earth seeing a smartphone. I started reading the manual and connected it to the TV. Then, I saw the computer prompt. I just started coding and didn’t leave my room that summer. I was coding at the time the internet didn’t exist. The storage was a music cassette drive. No internet. The microprocessor was like 7.16MHz. 64KB of RAM. I got really into it. I released my software on a site similar to Craigslist and people were buying my software. It was amazing! People were buying my software. After selling about 50 - 100 copies, I started getting calls from people complaining about the software. I realized customer support is hard so I stopped selling it. But, that changed my mind from becoming a doctor to studying Computer Science. I studied like crazy ever since all the way to my PhD.
Benny: How did you get interested in robotics?
Paolo: It happened at the university. We had to build something hands-on. When we were learning about electronics, we had to build an amplifier. In the second semester, I had to build a computer from scratch. We had to design it, build it, and make sure it works. For my Master degree project, I was walking from lab to lab to see what project is available. Then, I ran into a professor doing robotics. I was fascinated and did my master with him. I started coding the robot and, for 6 months, I didn’t leave the lab. When I finished my Master's, the professor asked me if I wanted to do a Ph.D. immediately. I said, “No, this is too hard and I want to earn some money.” He replied back, “Take your summer off and think about it.” After the summer, I decided that I want to pursue a Ph.D. and I did it with him in robotics.
Benny: What was your specialty in robotics?
Paolo: Computer vision to analyze tasks and navigate through space - how do you know where you are and where you want to go? That was my core specialty in my Ph.D.
Benny: How did you go from that to the robots you are building currently, which is supposed to be more human-like?
Paolo: My previous company invented this technology called VSLAM, a visual navigation technology that allows the robot to map the environment using the camera so it knows where it is. We sold the company to iRobot. Now, iRobot is using this technology in their products to clean your floors. It knows how to go from one room to another. It can cover every square inch of your floor because the [vacuum robots] know where they are. I spent about 10 years developing this technology in the consumer robotics area. The only successful robotic product in the market is vacuum cleaner. I have a lot of passion for this technology. But, I never imagined spending my career on floor cleaning after my Ph.D. That wasn’t exciting in terms of impact on society. But, it was exciting in terms of the technology itself. My original vision for my last company, Evolution Robotics, is to provide elderly care. But, we never got there. Then, after selling the company to iRobot, I decided to continue on the vision in building robots that have impacts on people’s lives, which is intersecting robotics with healthcare. That’s how we got to Embodied today where we are helping kids with neurodevelopment disorders like autism. Long term, we want to help elderly care as well. Also, technology has come a long way. This wouldn’t have been possible 10 years ago.
Part 5: Technological change that enables Embodied
Benny: Tell us about the changes in technology that makes it possible now.
Paolo: Speech recognition and natural language understanding. Ten years ago, you could barely say one command and have the machine recognize it. Today, you have devices like smart speakers at home where you can say sentences. It is very accurate. Remember we had speech recognition technology in cars where we gave it commands. By the time you input the address in the navigation system, you have reached your destination. Now, that technology has advanced. Thanks to deep learning.
The other advancement is computer vision - the ability to recognize things in the environment. Especially in our product, we have to recognize the child. We have to recognize their eye glaze where they are looking. We can recognize their facial expression. Ten years ago, we had to hard code everything. For the iRobot vacuums, it took us many years because we had to hand code every single scenario that could go wrong. Whereas nowadays, you can download databases of examples with images that are labelled. There are machine learning libraries that are available publicly from any company from Facebook, Google, Amazon, to Microsoft. The algorithm trains itself on how to recognize facial expression. Deep learning also had a huge impact on this.
Thanks to Moore’s law. The cost of computing continues to come down. My first computer in the late 80s had a microprocessor that was 7.16MHz with 64KB of RAM. Today, you can buy a computer for a fraction of that price with a million times that processing power.
The 3rd important change: ten years ago, if I walk into a VC meeting and tell them that I am working on robotics, that would be the end of the meeting. Today, there is significant investment into the robotics area. Even if you take out self-driving cars, there is billions of dollars invested into robotics.
The 4th element is talents. Before, it was hard to find experts in computing vision. Now, it might be hard to hire them because it is very competitive. But, it is possible to find them now.
For today's standard, what we are doing is extremely ambitious. Ten years ago, it was impossible.
No one has done this before. We are putting together many different technologies to create an amazing experience for the child. If we want to succeed, technology has to be seamless like an experience, not technology.
The thing that pushed the technology forward is the smartphone market. The smartphone market continues to push the boundaries of many areas from deep learning, cost, power, performance of computation, resolution of camera, etc. That has helped to advance the field tremendously.
Part 6: Being a compassionate technology company
Benny: It is very interesting to have this conversation with you about technology and compassion since you are building robots to serve people. Compassion needs to be embedded in the culture of the company.
Paolo: Absolutely. We want to serve families with kids with developmental challenges. These are kids who are being discriminated against everyday because they are different. They behave differently but they are not stupid. They might not be able to express their emotions.They might not be able to communicate or socialize their emotions. It is heartbreaking when your child comes home devastated when other kids treat them badly because they are different.
One of the things we do at our company is to hire people with autism as well because we notice they are very good at certain things. We like to hire them and empower them on this mission. This mission attracts people of similar backgrounds because they understand the pain and they want to help. That is part of our culture for sure. We have this whole notion of inclusivity and diversity at a very high level.
When it comes to technology, we do not believe that a robot should replace the needs for interacting with other people.
“We do not want to build robots that replace the need for humans.”
We are building robots that’s going to be a tool that allows your child to exercise these skills so they can be successful. It is like if you want to compete athletically, you would go to the gym and use exercise machines to get stronger. The point is not to use the machines so you don’t interact with people. The robot is a tool that allows the child to exercise in a safe environment so they can strengthen their skills for interacting with other kids successfully and to help them overcome the challenges that they were born with.
Benny: This means you don’t want robots to be with kids all the time.
Paolo: It is part of our design. We don’t want robots to be like an addictive video game that steals the attention of the child all the time because that actually worsen the problem. If kids don’t spend time socializing with other people, they get worse in socialization. The same is true for normal kids.
As an engineer myself, many engineers don’t spend enough time with people and they become awkward around people. Part of the reason is because they spend all their time in front of their computer. They don’t go out and socialize. If you look at MBA students, they are charming and outspoken. They are confident in front of a group and give presentations because these are the skills they have developed over time.
If you isolate the child with a machine, they would get worse and worse in their social skills. You want to encourage the child to go out to the real world and practice the skills you have trained them on. Our robots would shut down after so many interactions, give the child a mission to go out to the real world, practice, and report back.
Benny: In some way, your robots will be helpful for engineers and scientists as well.
Paolo: The big thesis of the company is that for so many years, we have learned to adapt ourselves to technology because of the limitations of technology. We have keyboards, mouse, etc. This is not a natural way to interact.
Now, we have the opportunity for the machines to adapt to humans. Let machines learn to interact with us using our body language, our verbal language, eye contacts, and emotions.
If we succeed at doing this, then every computing device and every machine should be able to interact with you the same way you interact with other people. It removes the limitation due to keyboards and screens. Everyone will use social skills with machines as well. It will be part of your life. I am 100% certain this is inevitable whether we are the company that gets there or not. This will happen. It’s only a question of time.
Benny: By adapting machines to people instead of people adapting to machines, I can see why this is a tough technical problem.
Paolo: Robotics is known to be a hard field because it is multi-disciplinary. In our company, we have mechanical engineers, electrical engineers, software engineers, system engineers, etc. In addition, we have taken diversity to another level. Since the early days, we have Craig, our Chief Creative Officer. Craig comes from entertainment. He designs characters that people would fall in love with. We have a whole design team that design experience that’s going to be fun and endearing. We also have child development experts because they understand how the child mind develops in early development.
Paolo: Getting back to how we get everyone to understand each other and the inter dynamics. In this company, I realize it is much harder when you have people with diverse backgrounds. In engineering, at least we have a common language. When you bring in designers, they have a different language and different mindset. It has taken some time to develop the vocabulary so everyone can understand each other.
Paolo: At the end of the day, it starts at the top of the company. If the top doesn’t lead with the values, people will not listen to what you say. They would model what you do.
Benny: Without a question, compassion is even more important for you given your product and your vision.
Paolo: I am building this company because I believe in it, not because I have to do it. I am building it based on the core values that I have developed over my life. For me, it comes very naturally.
“Now, we have the opportunity for machines to adapt to humans instead of humans adapting to machines. ”