In March, SNU launched the ‘System Semiconductor Engineering for AI’ combined major. Samsung Electronics and SK Hynix will support the management of the combined major based on the agreement signed with SNU.
Hyuk-Jae Lee, Chair of SNU ECE, who supervised the launch of the combined major, met with IT Chosun to explain the major’s curriculum and system.
SNU ECE Professor Hyuk-Jae Lee
―’Please elaborate on the ‘System Semiconductor Engineering for AI combined major’.
Students who have finished two or more regular semesters of the undergraduate program can apply regardless of their department. The program will be run as a double major, in parallel with the students’ original major. The Department of ECE has overseen the establishment of the major and will manage the program starting this semester by selecting 80 students.
The aim is to nurture professionals for the development of AI semiconductors that are used for various applications and services such as AI, big data, IoT, and autonomous vehicles. We will provide education for not only semiconductor circuit and system design but also applications and software development. We hope to train talents knowledgeable in both theory and practice through experimental and practical education.
―There are quite a few other universities that have launched semiconductor majors based on agreements with companies. What is the difference?
Ours is not a contract-based department. Other universities are creating contract-based departments where students are promised jobs in the companies after graduation, whereas our students are not.
Some suggested establishing a contract-based department, but they were outnumbered by those who believe that it is not the role of universities to train talent for specific companies. Instead, we will have internship programs, lectures by experts, and scholarships in connection with companies.
―What do you hope to achieve through the combined major of Semiconductor System Engineering for AI?
First of all, it will help resolve the scarcity in professionals.
Out of approximately 160 undergraduate students in SNU ECE, only 20~30 students actually graduate after receiving specialized education in semiconductors. The situation is similar for other universities.
With the impending era of AI semiconductors, many experts with professional insight encompassing memory and system semiconductors should be produced. There should be a sufficient number of experts with Ph.Ds and postdoctoral researchers. More experts can now be trained through the combined major course.
Second, we expect synergy effects from students who have professional knowledge in other majors.
For instance, if a student knowledgeable in automotive engineering has a dual major in AI semiconductors, he or she will be able to develop innovative semiconductors for autonomous vehicles. A student who handles data mining may make improvements to the semiconductor by using the data accumulated with semiconductor experiments. During projects, they would be able to exchange knowledge related to their respective majors.
―I heard that there is a new practical training course developed for a balance in theoretical and practical education.
That is true. There are three subjects, two of which must be completed for graduation. Their names are AI System Design Project (AI 시스템 설계 프로젝트), AI Semiconductor Circuit Design Project (AI 반도체 회로 설계 프로젝트), and AI Semiconductor Device Design Project (AI 반도체 소자 설계 프로젝트).
In the AI System Design Project course, students train in the design and development of imbedded systems. They practice designing and applying applications, system architecture, system software, and microprocessors for various AI systems such as IoT and autonomous vehicles.
For the AI Semiconductor Circuit Design Project, students will have hands-on practice in designing and implementing hardware circuits and architecture for system semiconductor development. Testing of the designed circuit will be performed through simulation, FPGA (Field Programmable Gate Array) verification, and chip manufacturing and testing.
The AI Semiconductor Device Design Project course will cover the entire or partial unit processes required for CMOS manufacturing, including ion implantation, CVD (Chemical Vapor Deposition), oxidation, dry etching, and metallization. The purpose is to provide students with the experience of semiconductor device development and verification.
―System Semiconductor Engineering for AI is a two year course. Is this enough to produce professional talents?
Of course, it is insufficient. Students are taught only the basics during those two years. In addition to the regular courses, we hope to offer short-term courses during break and provide training in semiconductor tools to complement the regular courses.
The combined major is an opportunity to produce additional experts in the semiconductor field apart from majors. To advance into the AI era, the memory field is important, but the system semiconductor field must also be reinforced. There is a great shortage in experts. For example, on the one hand, Taiwan’s foundry company TSMC’s design house GUC (Global Unichip Corporation) has approximately 700 workers. On the other hand, the total number of Korea’s design house experts is about 600.
The solution is to cultivate a large number of people with expertise. There is also a lack of professors to teach these people. To this end, it is necessary to secure professor-level experts by intensive support for the postdoctoral researcher system.
I think that the first step is the System Semiconductor Engineering for AI course. I hope that students who are attracted to the semiconductor field through this major will continue on to the master's and doctoral programs to lead the next generation semiconductor field.
Translated by: Jee Hyun Lee, English Editor of Department of Electrical and Computer Engineering, email@example.com