Implementing a Retrieval-Augmented Generation (RAG) Chat Model for Enhanced Software Support

Félév: 2024-2025 tavasz

Kategória: Szoftver

Téma leírása

This project aims to develop and integrate a Retrieval-Augmented Generation (RAG) to improve software support processes, specifically for responding to support queries. The model will combine retrieval-based and generative AI techniques to provide relevant, context-aware answers to support teams, enhancing efficiency and reducing response time. The model will retrieve similar past cases from Jira and generate AI-powered responses to assist in resolving current support queries. Additionally, it will provide support agents with relevant solution recommendations based on historical Jira tickets, enabling them to resolve queries more.

 

Tasks to be performed by the student will include:

• Design the architecture of the RAG chat model and its integration with Jira support tickets.
• Define the query processing pipeline and retrieval mechanism.
• Gather and preprocess Jira support tickets, or other relevant resources and format the collected data for retrieval and generation, ensuring structured tokenization.
• Implement the RAG model and train it on the processed data to generate relevant responses.
• Evaluate the model performance using accuracy, relevance, and response latency metrics

Külső partner: Joao Cavalcanti, Software Engineer at Ericsson Hungary

Maximális létszám: 1 fő