🤖 HRChacha: AI-Powered Hiring Assistant

HR Chacha is an AI-powered hiring assistant chatbot built using Streamlit, LLaMA 3.3 70B, MongoDB and deployed with AWS ECR + EC2. It simulates an intelligent recruiter that collects candidate details, asks technical questions based on their skills, and formats their data into a clean JSON format for HR/admins to evaluate.

Live link: (AWS Live Deployment Paused) Streamlit -> https://hrchacha-bmerkgxmqzeiwvlkvjccwt.streamlit.app//

📌 Project Overview

HR Chacha is a conversational AI assistant built for recruitment automation. It:

• Greets candidates professionally
• Collects basic candidate info (name, email, experience, location, tech stack)
• Dynamically generates tech-specific questions (e.g., Python, Django)
• Accepts candidate answers, stores data in MongoDB
• Outputs final structured data for HR teams to assess

---

Screenshots

---

🚀 Installation Instructions

1. Clone the repo

git clone https://github.com/Shristirajpoot/HRChacha.git

cd HRChacha

2. Setup environment

python -m venv venv
source venv/bin/activate  # Windows: venv\Scripts\activate
pip install -r requirements.txt

3. Environment Variables

Create a .env file:

MONGO_URI=mongodb+srv://<user>:<pass>@cluster.mongodb.net/
TOGETHER_API_KEY=your_together_api_key_here

4. Run the app

streamlit run app.py

---

📖 Usage Guide

1. Open your browser to http://localhost:8501
2. Start chatting! HR Chacha will ask for your name, email, experience, and skills
3. Based on your tech stack, HR Chacha will generate 3-5 relevant questions
4. Once all answers are received, your structured JSON will be shown and saved to MongoDB

---

⚙️ Technical Details

🧱 Tech Stack

• Frontend: Streamlit
• Backend: Python
• LLM API: Together AI (LLaMA 3.3 70B Instruct Turbo)
• Database: MongoDB Atlas
• Deploymet: Github Workflow, Runners, Docker, AWS ECR, AWS EC2
• Prompt Management: Custom, static system prompt with extract logic

📦 Libraries Used

• streamlit
• pymongo
• python-dotenv
• together (for API access to LLaMA)

🧠 LLM Architecture

• Uses chat history and a persistent system role
• Messages streamed with cursor-style output
• Structured prompt instructs LLM to collect & serialize user data

---

🎯 Prompt Design

Prompting strategy follows best practices:

• Starts with a clear role definition
• Instructionally structured in numbered format
• Uses trigger phrase USER_DATA to tag final JSON output
• Prompts candidate to answer tech questions with number and answer pairing
• Handles irrelevant input with fallback behavior

See the full prompt in hrchacha/prompts

---

📦 CI/CD & Deployment

• Docker: for containerized development and deployment
• GitHub Actions: to automate test and deployment workflows
• AWS EC2 / ECR: for cloud hosting + storing image

graph TD;
  Dev-->GitHub;
  GitHub-->CI/CD;
  CI/CD-->Docker;
  Docker-->AWS;
  AWS-->Production;

Loading

---

👨‍💻 Author

• 👤 Shristi Rajpoot Pre-Final Year IT | ML + Fullstack | Building real-world solutions