In his talk, Gökhan provided insights into the local Large Language Models (LLM) developed at our institute, which support our cleanroom team by reducing interruptions in semiconductor manufacturing.

This is how the LLM system contributes to an overall increased production efficiency:

• Fast analysis of technical documents and historical records to provide guided troubleshooting support and efficient problem resolution
• Secure handling of sensitive and confidential data with locally running LLMs without requiring an internet connection
• Continuously learning system based on real-life experience by resolving real-world problems and recording them within the system

The integration of knowledge accumulated over many years by our institution and experts, particularly in the cleanroom domain, transforms the system from a simple AI chatbot into an experienced "colleague" working alongside you. 

If you are intrigued by this novel approach to smart semiconductor processing, you're welcome to contact Gökhan Kayhan and Martin Schellenberger, our Data Analytics group manager.