LLM based Chat Agent for Enterprise Application Development

Every company has its own way of doing things, and IT systems are no exception. Some companies use VMs with Ansible for deployment, while others prefer K8S or serverless. This can make it hard for product teams to keep up with the latest standards, and they may end up creating their own shadow IT systems. Even if they do know the company standard, they may not have the skills to deploy applications on it. This means they have to rely on the central platform team, which can create a lot of work for them.

This project aims to create an LLM-based chat agent that can help product teams with their IT needs. The agent will be able to:

• Identify the solution that the team needs
• Ask clarifying questions to get a better understanding of the team's needs like size of the application, access pattern (internal/public). If billing is involved, which department to determine the right cluster to use.
• Create a pricing estimation
• Generate example infrastructure code (Terraform) with a CI/CD pipeline (Github) for deployment

This is a challenging project, as company IT systems can be complex. Students will focus on at least two different application deployments, one for a public web server and one for an internal API, ideally each having a database backend. The pipeline tooling can be GitHub, and the infrastructure can be an existing K8S cluster or a serverless infrastructure like Cloud Run or AWS Lambda.