# Projects

## 1. NeuraChat AI Chatbot (2024)
**Tech:** Python, LangChain, LLMs, Streamlit, OpenRouter
**GitHub:** github.com/Codeabhi096

- Built a context-aware conversational AI chatbot using LLMs with multi-language support, prompt optimization, and session memory
- Achieved 40% improvement in response relevance through advanced prompt engineering
- Integrated OpenRouter for LLM API routing across 5+ model providers
- Reduced API latency by 35% and enabled seamless switching between OpenAI and open-source LLMs

---

## 2. Disease Prediction App (2024)
**Tech:** Python, Machine Learning, Deep Learning, CSS, Streamlit
**GitHub:** github.com/Codeabhi096/DiseasePredictionApp
**Stars:** 5 ⭐

- Built a machine learning-powered web application for early detection of Diabetes and Skin Diseases
- Leveraged deep learning and classification algorithms for accurate predictions
- Users can input medical data to get instant health risk assessment
- Designed clean UI for easy use by non-technical users

---

## 3. AI Voice & Gesture Control System (2024)
**Tech:** Python, OpenCV, Deep Learning, NLP

- Developed a real-time Human-Computer Interaction (HCI) system combining computer vision and NLP
- Achieved 92% gesture recognition accuracy with under 100ms inference latency on live webcam input
- Combined voice commands and hand gestures for seamless system control

---

## 4. AI-Based Smart Examination System (2024)
**Tech:** Python, OpenCV, YOLO, Flask, NLP

- Designed an intelligent proctoring platform using YOLO object detection and NLP
- Detected 95%+ misconduct cases accurately in real-time
- Built multi-feed monitoring system with automated analytics reporting
- Helps institutions conduct fair online examinations

---

## 5. Diabetes Prediction Web App
**Tech:** Python, Streamlit, Machine Learning, Jupyter Notebook
**GitHub:** github.com/Codeabhi096/Diabetes_pridiction_webapp
**Stars:** 4 ⭐

- ML web app that predicts likelihood of diabetes based on health parameters
- Users input medical data like glucose levels, BMI, age for instant prediction
- Built with Streamlit for clean, interactive UI

---

## 6. California Housing Price Prediction
**Tech:** Python, XGBoost, Streamlit, Jupyter Notebook
**GitHub:** github.com/Codeabhi096/California-Housing-Price-Prediction
**Stars:** 4 ⭐

- Predicts California housing prices using XGBoost Regressor
- Trained on California Housing dataset with high accuracy
- Deployed as interactive Streamlit web application

---

## 7. Algo Trading System
**Tech:** Python
**GitHub:** github.com/Codeabhi096/Algo-Trading-System-
**Stars:** 4 ⭐

- Python-based mini algorithmic trading system
- Fetches daily stock data for multiple NIFTY 50 stocks
- Applies RSI + Moving Average crossover trading strategy
- Logs trade signals and performance automatically

---

## 8. Customer Churn Prediction
**Tech:** Python, Jupyter Notebook, Machine Learning
**GitHub:** github.com/Codeabhi096/customer-churn-prediction
**Stars:** 4 ⭐

- Predicts customer churn for businesses using classification algorithms
- EDA-driven feature engineering for improved model performance

---

## 9. AI Brainstorm Tool
**Tech:** Python
**GitHub:** github.com/Codeabhi096/ai_brain_storm
**Stars:** 4 ⭐

- AI-powered brainstorming assistant
- Helps generate ideas and creative solutions using LLMs