Categories
Misc

Edge Computing in Ethiopia – A Quest for an AI Solution

This is a guest submitted post by Natnael Kebede, Co-founder and Chief NERD at New Era Research and Development Center It was one year ago in a random conversation that a friend told me about a piece of hardware in excitement. At that point I never imagined how that conversation would have the potential to … Continued

This is a guest submitted post by Natnael Kebede, Co-founder and Chief NERD at New Era Research and Development Center

It was one year ago in a random conversation that a friend told me about a piece of hardware in excitement. At that point I never imagined how that conversation would have the potential to impact my life. My name is Natnael Kebede, the Co-founder and Chief NERD at New Era Research and Development (NERD) Center in Ethiopia. NERD is a center that provides a hacker space, educational content, and research for Ethiopian youth to create a better Ethiopia Africa and World

That piece of hardware was the Jetson Nano. I went over to my house and started researching about the Jetson Nano that night. From that point onward, I could not stop following up and researching about the NVIDIA edge computing concept. This is a story about how a single conversation helped me build a career and a community around an idea.

The conversation we had with my friend was about edge computing disrupting the new AI-development environment. For a country like Ethiopia, AI is usually considered a luxury than a necessity. This is due to the perception of universities, investors, startups and the government have about AI. All of them think about expensive high-performance computers and lack of experienced professionals in the field. Very soon, I realized how the edge computing solution could be a peaceful weapon to change the attitude towards AI in Ethiopia. Me and my partner decided to buy the Jetson Nano and put our name on the map.

The learning process was much easier and efficient when we started experimenting with the Jetson hands-on. It was coincidentally by that time we were invited to the first Ethiopian AI Summit to showcase any project related to AI. We decided to go for building an edge solution. Our research team side of the problem was to build a system that reads a streamed video, and use the Jetson to identify any desired inference at the edge. The application could range from counting cars in a connected traffic junctions and detecting license plates, to an agricultural solution where any flying or ground vehicle feeding a video of a farm to detect plantation health issues/counting fruits. We started off with the traffic management system. We organized a team of three engineering interns and myself to build a prototype in less than eight weeks.

The NERD team behind the traffic management system.

We got the whole thing figured out after going through a lot of reading. Finding answers was not the easiest thing as it was rare to find publications on the Nano. Meanwhile, a friend gave our contact to one of the NVIDIA Emerging Chapters Program leads, and we had a conversation about the program. It was the best timing. Although we wish we had known about the program earlier to access the DLI (NVIDIA Deep Learning Institute) courses, we kept going with a hope that the program will enable future projects with hands-on experience and technical training. The project was finally presented at the Summit and we had the chance to pitch the idea of edge computing to the Prime Minster of Ethiopia His Excellency Dr. Abiy Ahmed. We received promising feedback from him in regards to taking the project forward into implementation on the years to come.

Presenting the smart traffic light control system to Prime Minister Dr. Abiy Ahmed at the Ethiopian AI Summit

After the Summit we knew we needed more talent to recruit and inspire. We launched a short-term training on the Jetson Nano. It took our team eight weeks to build the traffic management system. We made sure the training takes the same duration of time for the students to build a final project at the end. We value open-sourcing the project. The output from these trainings are projects to be open-source so we can build our community bigger. Since we have limited resources, we only provided 25 students in three groups. Currently, we registered 13 students and the first batch is currently taking trainings. We are using the free DLI courses we were granted as part of the Emerging Chapters Program as guideline to our curriculum. We will soon provide them with free course vouchers after filtering consistent training participants.

First group of Mechatronics at the Jetson Nano training

Our goal is to create a community of enthusiasts, hobbyist, engineers and developers passionate about edge computing solutions in Agriculture and Smart City projects. We do this by consistently engaging with the tech community in Addis Ababa, Ethiopia. We organized an open seminar last week where we showcased our projects and also talked about edge computing. We were able to inspire more students for our program. Our team members are all enjoying the free DLI courses and we will be coming up with something much larger very soon. I want to personally thank the NVIDIA Emerging Chapters Program for all the support and resources. On behalf of our team, we are grateful to be part of the program. Very soon we would love to present our work for other partners!

The AI-based traffic control system is now on GitHub.

To learn more about NERD and to stay up-to-date, follow us on Instagram @nerds_center.

Leave a Reply

Your email address will not be published. Required fields are marked *