Gen AI Hackathon

This a chatbot built to shortlist candidates for a job interview. The goal is lessen the burdern on the HR team by automating the process of shortlisting candidates.

The chatbot uses Generative AI to rate the candidates based on their responses to the questions asked. Depending on the total score, the candidate is allowed to schedule an interview or not.

How it works:

• Every time a candidate enters a response to a question, the response is sent to the LLM model via the backend.
• A prompt is generated using the response and the predefined prompt template. Following is the prompt template:

template = """
You are an AI designed to evaluate the quality of a user's answer based on the question provided for a {subject} interview.

Question: "{question}"

User's answer: "{user_answer}"

Rate the user's answer on a scale from 1 to 10 based on its relevance, accuracy, completeness (with 1 being completely irrelevant or incorrect, and 10 being a perfect answer) and provide reason for your rating in under 100 characters.
{format_instructions}
"""

Link

https://gen-ai-interview-bot.streamlit.app/

Architecture

The following architecture is only for the custom FastAPI and ReactJS app. The Streamlit app has a different architecture.

Components

• The frontend is built using ReactJS

• The backend is built using FastAPI

• The chatbot is built using meta-llama/Meta-Llama-3.1-8B-Instruct-Turbo LLM model.

• After the prompt is generated, the model generates a rating and a reason for the rating.

• The rating and the reason are then sent back to the frontend where the candidate can see the rating and the reason.

• The candidate can then schedule an interview if the rating is above a certain threshold.

Pre-requisites to run the app locally

• First create a free account on Together AI. This is required for api key

• Create a new file named .env in the backend folder and add the following content:

TOGETHER_AI_API_KEY=YOUR_API_KEY

FastAPI

To run the FastAPI app, fist update the requirements.txt file by uncommenting all the packages after the following line: # for fastapi

Then run the following commands:

cd backend
pip install -r requirements.txt
uvicorn app:app --reload

Frontend

If you want to run the Frontend app, you need to create a new file named .env in the frontend folder and add the following content:

VITE_API_URL=http://localhost:8000

To run the frontend app, run the following commands:

cd frontend
npm install
npm run dev

Streamlit App

Create a new file inside .streamlit folder named secrets.toml and add the following content:

TOGETHER_AI_API_KEY = "YOUR_API_KEY"

To run the streamlit app, run the following command:

cd backend
pip install -r requirements.txt
streamlit run streamlit.py