Inspiration
What it does
CallLLM is a project focused on analyzing customer support calls using open-weight large language models. The goal is to extract key insights from voice transcripts — such as sentiment, intent, urgency, objections, and patterns in customer behavior — and turn them into structured data for business decisions.
What inspired me to join this hackathon was the opportunity to explore OpenAI’s new OSS models in a real-world context, without relying on hosted APIs. I believe LLMs can fundamentally improve post-call analytics and automation in contact centers.
Through this project, I’ve learned how to design prompt pipelines for noisy and unstructured input like transcripts, and how to use open models locally for faster and more private inference. I’m especially excited about the opportunity to engage more with the OpenAI ecosystem, connect with others, and build tools that push the limits of local LLM workflows.
Log in or sign up for Devpost to join the conversation.