Function Calling - Connect LLMs to the Internet

Ashish Tiwari (~ashish187)





The integration of Large Language Models (LLMs) with the internet through function calling is a transformative step in the evolution of conversational AI. This talk will delve into the innovative technique of enabling LLMs to intelligently generate JSON objects that contain arguments for calling one or multiple functions. This capability allows seamless connections to various data sources and APIs, significantly enhancing the functionality and application of LLMs.

During this session, I will demonstrate the process of implementing function calling using ChatGPT and building a complete workflow in Python. The demonstration will include connecting multiple data source points, such as NLP to Elasticsearch queries, and integrating with external APIs and other public sources. Attendees will gain insights into the practical aspects of setting up such a system, the challenges involved, and the solutions to overcome them.


Having knowledge of ChatGPT APIs will be helpful.

Content URLs:

Speaker Info:

Ashish started his career in the email and software industry, working as a software engineer. He has a strong interest in technology and now serves as a DevRel Engineer at Elastic. In this role, he is committed to providing the community with innovative solutions using the Elastic stack. Ashish is also a former co-founder of a startup, where he acquired a wide range of skills in tech, marketing, and community engagement. His enthusiasm is especially strong in the areas of coding and technology experimentation. Recently, he has delved into exploring Generative AI.

Speaker Links:

Section: Artificial Intelligence and Machine Learning
Type: Talk
Target Audience: Intermediate
Last Updated: