Add README config

README.md

@@ -1,78 +1,11 @@
-# 🤖 Self-Correcting Data Validation Agent
-### Schema-Enforced, No-Hallucination Agentic Data Pipeline
-
-A production-style AI system that converts messy employee text into schema-valid JSON using:
-
-- LangGraph (state machine orchestration)  
-- OpenAI LLMs (structured extraction)  
-- Pydantic v2 (strict schema validation)  
-- Pandas (deterministic execution)  
-- Streamlit (interactive UI)
-
----
-
-## 🧠 Problem
-
-LLMs hallucinate missing fields and fabricate identifiers.
-
-This project enforces:
-
-- Strict schema validation  
-- Bounded self-correction retries  
-- Deterministic query execution  
-- Explicit rejection handling  
-- Zero fabrication of required fields  
-
 ---
-
-## 🏗 Architecture
-
-Messy Text  
-→ LLM Extraction  
-→ Schema Validation  
-→ Self-Correction Loop (if needed)  
-→ Final Valid JSON  
-
-Records missing required fields are rejected — never hallucinated.
-
+title: Self Correcting Data Validation Agent
+emoji: 🤖
+colorFrom: green
+colorTo: blue
+sdk: docker
+pinned: false
 ---
 
-## 🔁 Agent Flow
-
-extract → validate
-if fail → correct → validate → repeat
-finalize
-
-Retry attempts are limited and controlled.
-
----
-
-## 💬 Deterministic Query Engine
-
-1. LLM generates structured query plan  
-2. Pandas executes it  
-3. LLM summarizes computed results  
-
-No synthetic answers.
-
----
-
-## 🚀 Run Locally
-
-```bash
-git clone git@github.com:Siddhesh-Ai9797/self-correcting-data-validation-agent.git
-cd self-correcting-data-validation-agent
-
-python -m venv .venv
-source .venv/bin/activate
-pip install -r requirements.txt
-
-export OPENAI_API_KEY="your_key_here"
-streamlit run app.py
-
----
-
-# Stress Testing
-python -m src.eval.run_agent_suite
-
-
+# Self-Correcting Data Validation Agent
+Production-style LLM agent using LangGraph, Pydantic v2 and self-correction loops.