AI platform company PCN announced on the 24th that it has won a new project in the electronics and semiconductor sector under the Ministry of Trade, Industry and Energy's "2026 Manufacturing Tacit Knowledge-Based AI Model Development Project."
Manufacturing tacit knowledge refers to knowledge and judgment criteria acquired by skilled technicians through long-term experience but not documented or organized into data. The Ministry of Trade, Industry and Energy is promoting this project to convert such know-how into AI training data in response to the aging of skilled workers and the decline in youth workforce.
The Ministry of Trade, Industry and Energy will invest 48 billion won from this year's supplementary budget to select 30 manufacturing processes with high risk or severe labor shortages, supporting the construction of tacit knowledge data and AI model development.
The project involving PCN targets the semiconductor Yangsan testing process. System semiconductors undergo multiple tests at the wafer and package stages. Given the prevalence of small-batch, multi-variety production, it is a field where complete standardization and automation are difficult due to differing test conditions and judgment criteria for each product.
In particular, the process of determining the causes of defect patterns and follow-up measures has relied on the experience of skilled engineers. This project aims to structure these judgment criteria into data and develop and demonstrate an AI agent within nine months that supports everything from verifying the integrity of test results to interpreting anomalies, estimating causes, and providing handling guidelines.
PCN will be responsible for developing an integrated platform for AI training data and a knowledge search and reporting system. It will develop parsers to analyze STDF and CSV formats used in semiconductor testing data, and build functions such as similar case search based on Retrieval-Augmented Generation (RAG), dashboards, and automatic report generation.
PCN plans to utilize its own AI agent "OzAI" and the "Oasis" product line, which supports data collection, management, and utilization, for this project. Once the system is built, it will automate repetitive tasks such as data cleaning and report writing, allowing skilled technicians to focus on final judgment and inspection, according to the company.
Song Kwang-heon, CEO of PCN, said, "We will convert the judgment criteria of master craftsmen into data assets and develop them into an AI decision-support system that can be utilized across the entire organization," adding, "This will contribute to improving quality competitiveness at semiconductor manufacturing sites."
in TECH
