Meet Sebastian, our Computer Vision Engineer!
Learn about Sebastian's Life as a Computer Vision Engineer!
Sebastian joined Sevensense in 2019, after completing his Master's degree in mechanical engineering at ETH Zurich. With his experience in engineering and programming, he knows how to come up with the best algorithms for our robots. He is a valued member of the Localization Team, where he helps our robots with one of their most important tasks: understand their own position in the environment where they operate.
When did you discover your interest in robotics?
I did not have an exact moment where I found my interest, but rather I have just always pursued what I enjoyed the most. My family is mostly working in business-related jobs, so I was not much exposed to engineering in my childhood. But already as a kid, I found it interesting to understand how things work, and why. Even simple questions such as “why does laundry dry when hanging it up?” used to spark my curiosity about what happens around me. That's why I decided to study mechanical engineering: It helped me to understand how the world works, from a more applied perspective.
My interest in programming was raised in some of the computer science and robotics classes. The fact that one could interpret or manipulate the real world by writing instructions on the computer was fascinating to me - and still is.
What fascinates you most about robotics?
It’s like building a small world inside our reality - and then interact with it. The challenges one faces are very diverse and one gets to learn about new things almost daily. Another thing I love about robotics is that you are getting to the bottom of things. It is like trying to solve a puzzle and you have to make every detail match.
What exactly do you help robots do?
I make robots understand their own position during the time they are moving. This is a crucial task because robots can not navigate autonomously from one place to another, without knowing their own location.
At Sevensense, we have a whole localization team that works only on optimizing the self-localization of the robot. Hereby, we have two different goals: online and offline estimation. Offline estimators update and maintain the general map of the robot, while online estimators make the robot understand where it is right now in the moment the robot is running. Depending on the priorities and projects, I get to work on both parts, which I find very interesting.
How does a normal day look like for you, also in regard to Corona?
These days I work from home most of the time, except for some of the days where I get to test my algorithms on the robot. A typical day usually starts with the stand-up meeting, where we discuss the priorities and goals of the day with my team. Apart from that, I get to focus a lot on my work, with relatively few distractions. I guess working in a startup environment keeps the general overhead low, meaning that I have a lot of time to work on algorithms and solving important problems.
I am also part of the code culture team, which has the purpose of educating other programmers at Sevensense and developing coding standards that are beneficial for the developers as well as the company.
In general, I try to take regular breaks to clear my thoughts. When we were still going to the office, it helped to have a chat or laugh with my colleagues, but during home office times, this of course doesn’t happen automatically. Another thing that is very important to me is my hobby: dancing. After work, I always try to go training. This helps me to concentrate and come up with new ideas.
What is your biggest challenge at work?
The work we do at Sevensense is very complex since robotics systems are fairly complicated. We have different software and hardware components that all have to work together nicely. For me, the biggest challenge is to create something that works not only on its own but in the end, also together with the other components. This can be hard because sometimes it seems like we are working well on one part alone, but when combining multiple modules together we see that the result is not as expected. For me, this is also the interesting part about robotics - the challenge of building something big and powerful from well-designed smaller components.
What do you like most about your work as a computer vision engineer?
Some of the problems we are working on are fairly challenging. Therefore, when the moment comes where we find a fitting solution for it, and the solution works - that’s a great feeling! I also enjoy the work that comes when we have identified the best algorithm to tackle a problem. In the end, it’s not only about solving the problem but solving it in the best possible way. The code one leaves behind needs to be well readable, maintainable, and tested for whoever will need to work with it again, therefore, that’s another challenge on top of choosing the right algorithm.
We work together on a very ambitious goal: to make the world a more efficient, sustainable, and safe place by enabling mobile robots to navigate where they are needed the most.
What is your favorite thing about Sevensense?
That we are a group of people that are extremely passionate and motivated about what we do. We work together on a very ambitious goal: to make the world a more efficient, sustainable, and safe place by enabling mobile robots to navigate where they are needed the most.
We are all trying to bring this forward, and for me, this is very inspiring and allows me to learn a lot.
You have a very interesting hobby: you dance ballet. How does this complement your interest in robotics?
For me, ballet is like the math of dancing. It is very systematic, structured, and defined on what the goal is, meaning that there is a strong feeling of what is right or wrong, contrary to many other dances. I think this is the connection to my other interests. Still, what I like the most, is to learn how to form shapes with my body and move muscles in a way I did not even know was possible. In the end, you always create something new, and this is also what amazes me about robotics.