ReUnion ✦

"To bring together again that which has drifted apart—talent and opportunity—through radical transparency."

ReUnion is an intelligent Career Assistance Platform built for SCDS TechFest 2026. It bridges the widening chasm between evolving employer requirements and static candidate profiles by transforming fragmented job data into personalized, actionable upskilling roadmaps.

---

📖 The Philosophy

In the modern tech ecosystem, a lack of transparency leads to "hiring scope creep," where job requirements silently drift away from the skills candidates actually possess.

ReUnion derives its name from two core pillars:

• Re (Again): Acknowledging that employment is a continuous cycle of re-evaluation and re-skilling, not a one-time transaction.
• Union (Together): Bringing disparate entities—"Job Requirements" and "Candidate Profiles"—back into a cohesive whole.

Like the volcanic island of La Réunion that constantly reshapes itself, this project aims to be the stable ground in the chaotic ocean of recruitment data.

---

🚀 Key Features

1. Centralized Intelligence (The "Brain")

• Multi-Source Ingestion: Aggregates data from diverse sources including official CSV datasets (MyCareersFuture, Glassdoor) and unstructured PDFs (Problem Statements).
• Hybrid Search: Combines SQL-based hard filtering (Salary, Location) with Vector Semantic Search (RAG) to find hidden opportunities.

2. Actionable Roadmaps

• AI Gap Analysis: Uses Large Language Models (DeepSeek V3 / OpenAI) to analyze the specific "delta" between your resume and a target job.
• Skill Acquisition Plans: Generates linear, step-by-step learning paths.
• Local Integration: Automatically suggests relevant SkillsFuture courses and government subsidies.

3. Transparent Tracking

• Application Status: Track your journey from "New" to "Offer" directly within the roadmap interface.
• Tech Stack Filtering: Dynamically filter opportunities based on the specific technologies (e.g., Python, React, AWS) extracted from job descriptions.

---

🛠️ Tech Stack

Frontend

• Streamlit: Highly customized with a "Transcendant" Glassmorphism CSS theme.
• Dynamic UI: Real-time interaction with the backend database.

Backend

• FastAPI: Robust REST API serving the AI agent and data synchronization endpoints.
• SQLAlchemy & SQLite: Relational storage for structured job data and user profiles.

AI & Data

• LangChain: Orchestrates the reasoning chains for Gap Analysis.
• ChromaDB: Local Vector Store for semantic document retrieval.
• DeepSeek / OpenAI: Powering the cognitive reasoning engine.

---

📦 Installation & Setup

Prerequisites

• Python 3.9+
• An API Key (DeepSeek or OpenAI)

1. Clone & Install

git clone https://github.com/yourusername/SCDS-TechFest26.git
cd SCDS-TechFest26
python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

2. Environment Configuration

Create a .env file in the root directory:

# Choice 1: DeepSeek (Cost-Effective)
DEEPSEEK_API_KEY=your_key_here
LLM_MODEL="deepseek-ai/DeepSeek-V3"
LLM_API_BASE="https://api.siliconflow.cn/v1/"

# Choice 2: OpenAI (Stable)
OPENAI_API_KEY=your_key_here
EMBEDDING_MODEL="text-embedding-3-small"

3. Ingest Data (Build the Brain)

Initialize the database and vector store with the official datasets:

python3 ingest.py

---

🚦 Usage

Start the Backend

The backend serves the API and the Database connection.

python3 main.py
# Backend runs at http://localhost:8000

Start the Interface

Open a new terminal to launch the User Interface.

streamlit run web_app.py
# Frontend runs at http://localhost:8501

---

📄 License

Project created for SCDS TechFest 2026.