Graduates of the program will be able to accelerate computer vision and deep learning solutions within the Intel OpenVINO toolkit.
Intel and Udacity announced that their Intel Edge AI for IoT Developer Nanodegree program is open for enrollment. The collaboration, announced on Thursday, aims to train developers in deep learning and computer vision to help facilitate the deployment of artificial intelligence (AI) at the edge.
SEE: Special report: From cloud to edge: The next IT transformation (free PDF) (TechRepublic)
"We didn't build this program just for hobbyists; we built this for practitioners," said Alper Tekin, CPO at Udacity. "At the end of these courses, you have to write actual code that works in a production environment."
Focusing efforts at the edge
Internet of Things (IoT) and edge computing were areas Udacity had its sights set on because of the technology's popularity, as well as the skills gap that exists in the field. The company decided to build curriculums based on areas that have known gaps, Tekin said.
"The market for edge computing is booming. We see in our part of the business an opportunity north of $30 billion a year by 2024," said Matthew Formica, director of edge AI developer platform product marketing at Intel.
"As AI has been transforming the edge with its unique latency power and cost considerations, studies show that less than 20% of AI developers are trained for IoT," Formica said. "There's really a skillset gap there. We see that when we go out to customers and work with solutions builders--the technical skill isn't there."
Through the three courses, developers will become familiarized with Intel's OpenVINO software toolkit and be able to directly help launch AI and IoT projects at the edge.
The courses offered in the nanodegree
Students considering this program should have some familiarity with computer vision models and deep learning, Tekin said.
To earn the degree, developers must complete three program courses: Edge AI Fundamentals in OpenVINO, The Hardware for Computer Vision Deep Learning Application Deployments, and Optimization Techniques and Tools for Computer Vision Deep Learning Applications.
However, if students are only interested in one of the courses, they can still complete that single course without earning the nanodegree. The first Fundamentals course is free of charge, as a way for developers to get their feet wet, Formica said.
The other courses run $199 to $399 depending on promotional offers Udacity has at the time, Tekin said. Through all of the courses, students receive hands-on experience.
In the first course, "we have a bunch of pre-trained models for computer vision, and the project asks students to create a people counter app," Tekin said. "The students will investigate different models for person detection and come up with the best model for detecting and counting people. All this calculation will happen at the edge."
The second course focuses on hardware. "We're going to ask students to build their expertise in choosing the right set of hardware and considering the factors in selecting the hardware. CPU versus GPU, etc.," Tekin said.
"The project there will be across three different scenarios: Retail, manufacturing, and transportation," Tekin added. "We have scenarios where we're going to ask the students to come up with the trade-offs on what's the best hardware setup, and that will utilize the Intel DevCloud."
The last course is meant to optimize what the students have learned. "The students will be asked to, using the OpenVINO toolkit, to control their computer using their eye-gaze," Tekin said.
"The key technology building blocks that are driving AI at the edge is computer vision," Formica said. "It's all about pointing a camera at something and then inferring something about the real world from that camera. That's why the projects that [Tekin] talked about around people counting, gaze detection, and so forth, are so valuable."
Udacity and Intel specifically created the course with real-world situations in mind, so that students would be able to directly take their knowledge with them into their work.
"You see these in real-world situations like worker safety--making sure that workers are wearing the right protective gear," Formica said. "You can have a camera pointing at them. 'Yes, they have their helmet on. OK, we'll unlock the gate and let them go into the potentially dangerous manufacturing environment.'"
"One recent medical example I found interesting was how a company used our OpenVINO technology to build surgical hospital logistics robots that can carry medicine and supplies around the hospital without humans having to touch as many things," Formica said. "And they used OpenVINO and computer vision in that robot to be able to see its environment nearly six times faster than it was able to prior to applying this technology."
In November 2019, Intel and Udacity launched the course's free Fundamentals course to help narrow down scholarship applicants. Tekin said they only expected around 10,000 students to apply, but before long, more than 30,000 people applied to be considered for the scholarship.
The top 850 students who completed the Fundamentals course were awarded nanodegree scholarships on Thursday.
"By many estimates [the edge] is growing faster than the cloud did. And if you think about the amount of data that's processed in a factory, for example, it can be hundreds or even thousands of times as much data as ever gets sent to the cloud," Formica said.
"This is a critical skillset that students of all walks of life are going to need in the future," Formica added.
For more, check out How to capitalize on the edge analytics development life cycle on TechRepublic.
- Multicloud: A cheat sheet (TechRepublic)
- Hybrid cloud: A guide for IT pros (TechRepublic download)
- Serverless computing: A guide for IT leaders (TechRepublic Premium)
- Top cloud providers in 2020: AWS, Microsoft Azure, and Google Cloud, hybrid, SaaS players (ZDNet)
- Best cloud services for small businesses (CNET)
- Microsoft Office vs Google Docs Suite vs LibreOffice (Download.com)
- Cloud computing: More must-read coverage (TechRepublic on Flipboard)