Overview
This project was built for the Tools in Data Science course at IIT Madras. The goal was to design and deploy an LLM-based automation agent capable of handling structured tasks via a REST API interface.
What It Does
The agent accepts natural language instructions through a FastAPI endpoint and routes them to the appropriate tool or handler using an LLM backbone (OpenAI API). It can perform multi-step reasoning and execute actions based on the model’s output.
Tech Stack
- Python — core implementation language
- FastAPI — REST API server for agent interaction
- OpenAI API — LLM backbone for reasoning and instruction parsing