AIBIZ is a startup that develops a deep learning-based quality prediction system using sensor data from manufacturing plants. CEO Ha Seung-jae founded it in June 2020. After graduating from Seoul National University, Ha worked at Harim for more than 30 years. His stint at the group's holding company seven years ago became the impetus for starting his own business. "At the group holding company, I was in charge of guiding smart factory sites. That is when I became interested in the Fourth Industrial Revolution and big data. I judged it to be a market with sufficient potential, so I left the company and took on the challenge of starting a business. AIBIZ is a company striving to enter the rapidly growing 8-inch foundry semiconductor market." As the foundry semiconductor market booms, demand for semiconductor quality defect prediction is also increasing. In the existing semiconductor defect inspection market, reference values are set according to engineers' subjective judgment and people analyze them directly. Because basic statistical techniques are used in the analysis, productivity is low and tracing the causes of defects is difficult. There is demand to improve such defect prediction systems, but no suitable alternative exists. Ha said, "The defect prediction system is technically difficult and it is a field where securing data is also hard," adding, "It is hard to enter because there is also a shortage of experts with specialized knowledge in both artificial intelligence (AI) and semiconductors." To enter the market, AIBIZ built a team of experts with specialized knowledge in both AI and semiconductors. Both Ha, the co-founder, and the Chief Technology Officer (CTO) hold doctorates in IT engineering. Ha is a smart factory consultant and a data design expert with long field experience in processes and equipment. The CTO is also an AI solution development expert with a background in semiconductor R&D. Another strength is that the company carries out projects jointly with advisory professors who are AI experts and with industry-academia research labs. Ha said, "Because AIBIZ's team members are experts in the field, we were able to overcome the difficulty of the barriers to entry." AIBIZ developed a solution in which an AI algorithm automatically sets reference values from sensor data collected at semiconductor plants and provides engineers with graphs showing the causes of defects. Ha said, "Using this solution makes it easy to trace the causes of defects," adding, "It can raise national competitiveness by improving semiconductor yield and saving production costs." "Existing semiconductor defect prediction systems were limited to identifying defective products using machine vision on finished products. AIBIZ can predict defects during the semiconductor process, so it can improve yield. It tells engineers the causes of defects right away, enabling cost savings and yield improvement, and it is a one-stop service that can perform quality inspection not only in the front-end process but also in the back-end process." Compared with the technologies applied in most existing AI markets, AIBIZ has secured a technological competitive edge in terms of 'accuracy,' 'speed,' 'update method' and 'ease of use.' Ha said, "If you draw a positioning map of the semiconductor defect inspection market along two axes, 'price' and 'quality prediction method,' AIBIZ sits in the upper right, low cost and advance prediction, so it is highly competitive in the market compared with competitors." AIBIZ is currently providing its solution to 8-inch semiconductor fabs, where the effect of adopting AI is high because of aging equipment and diverse processes, and which continue to grow. Ha said, "Going forward, our goal is to provide the solution not only to the 8-inch semiconductor market but also to the 6-inch and 12-inch markets," adding, "With our differentiated AI technology and system, we plan to supply AIBIZ solutions to all manufacturing plants." Reporter Lee Jin-ho, Hankyung Job & Joy, jinho2323@hankyung.com