🛡️ Insight-Zero

The Autonomous Data Steward for Enterprise.

Insight-Zero is an autonomous, microservice-based data intelligence platform designed to replace the manual data extraction and anomaly detection loop. It features unsupervised statistical learning, PII sanitization, and a Context-Aware RAG (Retrieval-Augmented Generation) brain to not just detect anomalies, but explain why they happened.

---

🚀 Key Features

• Privacy-First Ingestion: Utilizes a custom NLP pipeline (Microsoft Presidio) to detect and redact Personally Identifiable Information (PII) before it ever reaches the AI analysis layer.
• Ensemble Statistical Engine: Combines parametric (Z-Score) and non-parametric (IQR) modeling to generate high-confidence anomaly scores without requiring massive labeled training datasets.
• Context-Aware RAG Brain: Integrates Pinecone Vector DB with Meta's Llama-3.1. Users can upload internal corporate PDFs, allowing the AI to diagnose root causes based on proprietary company context.
• Asynchronous Enterprise Ops: Powered by BullMQ and Redis, ensuring the platform scales to process massive datasets in the background without blocking the UI.
• Multi-Modal Reporting: Autonomously generates and exports boardroom-ready .pptx slides and high-resolution PDF incident reports.
• FinOps Telemetry: Built-in cost monitoring dashboard tracking processing latency and LLM token usage to prove operational ROI.

---

🏗️ System Architecture

This project is structured as a Monorepo utilizing a 3-Tier Microservice architecture:

1. Frontend (apps/web): Next.js 14, Tailwind CSS, Shadcn UI, Clerk Auth.
2. Orchestration Gateway (apps/api): Node.js, Express, BullMQ, Redis. (Handles secure routing, queuing, and background job polling).
3. Intelligence Engine (apps/engine): Python, FastAPI, Pandas, Presidio, Pinecone, Groq (Llama-3.1), Sentence-Transformers.

---

🛠️ Local Development Setup

1. Prerequisites

• Node.js (v18+)
• Python (3.11+)
• Redis (Local or Upstash Cloud)
• API Keys: Clerk, Pinecone, Groq

2. Installation

Clone the repository and install dependencies for all workspaces:

git clone [https://github.com/sdiv0503/insight-zero.git](https://github.com/sdiv0503/insight-zero.git)
cd insight-zero

# Install Node dependencies
npm install

# Setup Python Engine
cd apps/engine
python -m venv venv
source venv/bin/activate  # On Windows: .\venv\Scripts\activate
pip install -r requirements.txt

3. Environment Variables

Create .env files in the respective directories based on the .env.example templates.

• apps/web/.env.local: Clerk Keys
• apps/api/.env: Redis URL, Clerk Secret
• apps/engine/.env: Pinecone API Key, Groq API Key

4. Running the Platform

Start the microservices in separate terminals:

Terminal 1 (Next.js Frontend):

cd apps/web
npm run dev

Terminal 2 (Node.js Gateway):

cd apps/api
npx ts-node src/index.ts

Terminal 3 (Python AI Engine):

cd apps/engine
uvicorn main:app --reload --port 8000

👨‍💻 Author

Divyansh Sharma

GitHub: @sdiv0503

Email: sdivyansh0503@gmail.com

Developed as a Final Year Capstone Project — 2026

---