⚖️ EquiGrader AI

Fair, Explainable, & Stress-Free Technical Interview Prep

EquiGrader AI is an automated practice partner designed to level the playing field for technical interviews. By leveraging Local LLMs (Phi-3) and Speech Recognition (Whisper), it allows you to practice speaking naturally while receiving feedback based solely on your engineering logic—not your accent or grammar.

---

🚀 Why This Project Matters

• Truly Fair: Engineered prompts ignore grammatical slips or accents. If you understand the concept, you earn the points.
• No More Mystery Scores: Receive a detailed breakdown of rubric points hit and specific areas for improvement.
• Privacy by Design: Everything runs locally via Ollama. Your voice and data never leave your machine.

---

✨ Key Features

• Dual Practice Modes: Toggle between typing or speaking to simulate real-world pressure.
• Intelligent Reasoning: Powered by Phi-3, the system understands engineering logic rather than just scanning for keywords.
• Modern Interface: A clean, "Glassmorphism" styled UI built for focus.
• Professional Architecture: Decoupled FastAPI backend and Streamlit frontend.

---

🛠️ Tech Stack

Component	Technology
Frontend	Streamlit (Python)
Backend	FastAPI (Uvicorn Server)
The "Brain"	Ollama (Phi-3 Mini Model)
The "Ears"	OpenAI Whisper
Reliability	UiPath (Automated QA Testing)

---

⚙️ How to Get Started

Follow these steps to get EquiGrader AI running on your local machine.

1. Prerequisites

• Python 3.10+
• Ollama: Download here to run the AI model locally.

2. Setup the AI Model

Once Ollama is installed, open your terminal and pull the model:

ollama pull phi3

3. Clone the Repository

git clone [https://github.com/YourUsername/EquiGrader-AI.git](https://github.com/YourUsername/EquiGrader-AI.git)
cd EquiGrader-AI

4. Run the Backend (The Brain)

Open a new terminal window:

cd backend
pip install -r requirements.txt
python main.py

Note: Wait until you see: Uvicorn running on http://127.0.0.1:8000

5. Run the Frontend (The Interface)

Open a second terminal window:

cd frontend
pip install -r requirements.txt
streamlit run app.py

---

🔮 Roadmap & Future Vision

• RAG Integration: Allow the AI to "read" specific textbooks for targeted exam prep.

• Behavioral Analysis: Use Computer Vision to provide feedback on eye contact and confidence.

• Expanded Disciplines: Growing the question bank to include Mechanical, Civil, and Electrical Engineering.

---

Author

• Created with ❤️ and a lot of coffee by Hani Mohammad Kaif.