🚀 Hirelytics - AI Interview Mentorship & Analytics Platform

Welcome to Hirelytics, a next-generation AI-powered platform that transforms raw interview videography into highly structured, actionable mentorship. Built during the DevClash hackathon, Hirelytics moves beyond basic transcription to offer true "Real Mentor" insights—analyzing technical depth, acoustic vocal projection, and emotional composure in real-time.

✨ Key Features

• Upload-First Experience: Seamlessly ingest heavy video files with localized audio extraction to ensure fast, private processing.
• Deepgram Diarization: Pinpoint accuracy distinguishing between the Interviewer's questions and the Candidate's answers.
• Dual-LLM AI Engine:
  • Groq (Llama 3.1): Performs lightning-fast anomaly detection, instantly catching language switches or basic faults.
  • Google Gemini (1.5 Flash): Acts as the "Senior Technical Architect," grading structural logic, emotional composure (Fear, Anger, Hesitation), and generating detailed mentorship pipelines.
• The "Mentor Card": A stunning Glassmorphism UI that breaks down every critical moment into three actionable points: Diagnosis, Gold Standard (Correction), and a Growth Plan.
• Interactive Timeline Sync: Click a timeline event to instantly fast-forward the video and auto-scroll the transcript to the exact moment of failure or success.

🛠 Tech Stack

• Frontend: React + Vite, CSS3 (Custom Glassmorphism Tokens), Lucide React.
• Backend: Python + FastAPI.
• Audio/Video Processing: MoviePy, FFmpeg.
• Speech-to-Text: Deepgram Nova-2 (Diarized & Time-stamped).
• Core AI: Groq (Llama), Google Gemini.
• Database: Supabase (PostgreSQL).

🚀 Getting Started

Prerequisites

• Node.js (v18+)
• Python 3.9+
• Activated API Keys for: Deepgram, Groq, Gemini, and Supabase.

Environment Setup

Create a .env file in the /backend folder:

SUPABASE_URL=your_supabase_url
SUPABASE_KEY=your_supabase_service_role_key
DEEPGRAM_API_KEY=your_deepgram_api_key
GROQ_API_KEY=your_groq_api_key
GEMINI_API_KEY=your_gemini_api_key

Running the Platform

1. Start the Frontend:

cd frontend
npm install
npm run dev

2. Start the Backend: (Ensure you use the system Python with dependencies installed, bypassing local Conda loops if applicable)

cd backend
pip install -r requirements.txt
python3 -m uvicorn main:app --reload --host 127.0.0.1 --port 8000

📖 Deep Architecture

Curious about how the Dual-LLM Orchestration or the Acoustic Probability Engine works? Check out the Architecture Guide (ARCHITECTURE.md) for detailed data-flow diagrams and schema breakdowns.

---

Built with ❤️ by Team Rocket for DevClash.