photo: Marco Antonio Gutiérrez Giraldo, private archieve
Marco Antonio Gutiérrez Giraldo, aged 32 is one of the participants in NVIDIA Jetson™ Developer Challenge. Marco was responsible for the “Jetson AH” project. We’ve had a chance to ask him a few questions related to his participation in the contest, enjoy!
Marco, what are you up to now? Are you a student, are you employed?
I just finished my PhD in robotics and I am currently trying to setup a startup using AI to analyse network data.
Would you mind telling us about what led you to coming up with the idea for your project?
This project combines language and images to generate a multimodal AI system able to predict object locations. Back in my university I was doing Image processing so I had some skills in that area. Then I got a scholarship in the language technology department of A*STAR, a research center in Singapore, which allowed me to get the knowledge needed to combine both solutions for better results.
Why did you decide to take part in the NVIDIA Challenge?
When I was a PhD student I received a grant from NVIDIA that allowed me to work on AI for embedded devices (with Jetson TX1). Therefore, when I saw they were having a contest on the topic I thought it was the perfect match for my work. Plus I really enjoy working on robotics AI, so it was an easy decision!
Oh, that’s a very fortunate coincidence. And how long have you been interested in AI use? Have you built any system utilizing it before?
I was interested in AI since I started doing my PhD, during those years I worked on different kinds of robots with some sort of AI. Obviously this has been evolving as my research matured.
And when you were working on your project, how did it go? Any unexpected road bumps?
The project was quite tough, as always when you work with real robots and specially when you build everything from the ground up, like us did, it takes a lot of effort that actually involves people with a wide variety of skills. Dealing with constant hardware failures and bugs/mistakes is something you have to learn to deal with in a daily basis. However the final result is always quite gratifying.
What about your skills, did you develop any new ones?
The whole project matched my skills, robotics, AI and embedded systems. Being this an extension of works that I had developed in the past made it a project where I can fully develop my abilities and put them into practice.
That must mean a lot of work was a breeze for you! But there were definitely some downsides as well, could you tell us about both of those?
I actually spent more time than expected working on the submission materials, which is something that I do not personally enjoy much, although I understand is totally necessary, I rather be working on the real application than writing the whole process in paper, hehe.
Getting the really good results from the models and making sure they are an improvement over existing solutions took quite some time but finally we got a good gratifying outcome.
Did you have a chance to learn anything during the works on your hackathon project?
I've learned a lot, specially about AI on embedded systems, which is a field I had not experience with.
Now, if you don’t mind, let’s go into more general questions. What’s your take on the idea of hackathons?
I think hackathons are a good way to grow the spirit of building. They are very good when you are a student or learning a technology since they will help you to start building real things and will constitute a push on it. However hackathoners should also realize that building prototypes is not everything, there is also a need for the years of work that comes after to make real products.
And how often do you take part in challenges like this one?
I used to when I was studying, I think it's a good and fun way to challenge yourself and put into practice your new knowledge. As I am currently trying to build a full solution now I don't have that much time anymore.
Any plans for future hackathons?
I'm not sure, maybe if I have some time and the field and prizes seem interesting. Specially if it's a cool technology that I am starting to learn and I wanna expand my knowledge about it.
Well we certainly hope that you do! Do you think participation in hackathons will help you in the future?
I think taking hackathons in an early stage of your career on when learning new technologies is a good and fast way to boost your skills in the area. They did help me on my early years to obtain the abilities that I can now put into practice to build real products.
We couldn’t agree more. As for this one, did NVIDIA Challenge help you gain anything?
I wouldn't consider the NVIDIA Jetson Challenge competition as a typical hackathon itself as it was held for a longer time of period than usual hackathons (1-2 days). This allowed participants to actually develop solutions a bit more complete than mere prototypes in contrast to the actual outcomes of usual hackathons. I actually learned some good skills during the challenge that I can now apply into future projects. The project itself serves as a good base to continue doing research in the field. Oh, and ultimately I got a very interesting prize!
Let’s wrap this interview up with a piece of advice. Would you mind giving one to people who would like to participate in similar challenges in the future?
My advice to other competitors is first to try to look for those areas of your interest so what you learn can help you in the future. Then to try to think out of the box on innovative ideas but always with the practical mindset looking at it from the user point of view to see what is the real use of your solution. If your project and skills obtained from the hackathon are actually useful later on the prize is even better than actually winning the contest!
If you would like to participate in a similar hackathon visit our website with list of hackathons and challenges. Pick an event that suits your interests and get busy!