駿揚 彭

• Data Governance: Taking advantage of cloud computing, we use Azure Cloud and the distributed computing programming language PySpark combined with SQL for ETL development. We operate the adaptive factory data fusion horizontal expansion project in 12 factories and convert data sources into cloud data (integration promotion Plant-wide data management)
• Algorithm development:

1. Research and development of small amount and variety imbalanced classification algorithms: Conduct research on the small number of samples in each category of data due to product diversity and the imbalanced data in the categories.

2. Research and development of multi-label and multi-category classification algorithms: Classify multi-target data and perform multi-label prediction and classification at the same time.

3. Adaptive analysis process planning & development: The KPI indicator alarm automatically triggers the intelligent decision-making module to cascade algorithm analysis and provide machine parameter adjustment prescriptions, improving the analysis efficiency of integrated colleagues by 25%.

4. Mura identification & classification algorithm development: Pre-screening and manual determination of Mura by front-end inspection results can effectively reduce unnecessary processing and significantly reduce the manufacturing cost of new products.

5. Development of LLM intelligent analysis system: Strengthen LLM reasoning ability through knowledge graph to analyze various abnormal problems, emergency indexes, and provide corresponding feasible solutions, while collecting relevant feedback data to strengthen the knowledge base.