Workshop: Building LLM-Powered Applications on Commodity Computers

Abstract

This workshop presents various methods for optimizing large language models (LLMs) using open-source tools such as the Intel OpenVINO Toolkit, the Neural Network Compression Framework (NNCF), and Optimum Intel. The hands-on session enables participants to apply model optimization techniques and build domain-specific chatbots and Agentic AI applications on commodity computers such as laptops and desktop PCs based on Intel platforms. This workshop aligns with the scope of ICSIPA by empowering participants to build LLM-powered chatbots and agents on their own computers to augment the capabilities of their AI applications.