AI-Powered FARMIN-AI
Full-Stack Conversational AI System
Project Overview
Built a comprehensive conversational AI system designed to assist farmers with agricultural decisions. The platform provides intelligent recommendations for farming practices, crop selection, and problem-solving through an intuitive chat interface.
Key Achievements & Actions
- Developed a full-stack application using Streamlit for seamless user interaction.
- Trained and integrated machine learning models achieving reliable prediction accuracy.
- Deployed the application on Streamlit Cloud for accessible, farmer-friendly usage.
- Implemented natural language processing for understanding and routing complex farmer queries.
Technologies Used
- Frontend: Streamlit
- Backend & Logic: Python
- ML Frameworks: Hugging Face Transformers, Scikit-learn
- Data Processing: Pandas, NumPy
DWLR Data Monitoring Tool
Smart India Hackathon 2024
Project Overview
Designed and implemented a prototype for monitoring Digital Water Level Recording (DWLR) devices, capable of processing over 1,000 readings daily with advanced anomaly detection capabilities.
Key Achievements
- Processed 1,000+ DWLR readings per day with automated data validation pipelines.
- Implemented collaborative filtering between faulty detectors for enhanced accuracy.
- Developed anomaly detection algorithms to identify irregular water level patterns instantly.
- Created intuitive dashboards for real-time monitoring and reporting.
Telugu Speech Corpus
Research Internship at VISWAM.AI (IIITH)
Project Overview
During my internship at IIIT Hyderabad, I spearheaded the creation of a massive 80+ hour Telugu speech corpus to help train localized NLP and speech-to-text models.
Key Responsibilities
- Led meticulous data annotation for over 80 hours of raw Telugu audio.
- Performed rigorous model evaluation and validation to ensure dataset quality.
- Collaborated with AI researchers to optimize data pipelines for regional language models.
Conversational AI Chatbot
Hugging Face Transformers
Project Overview
Developed a custom conversational AI assistant leveraging pre-trained Large Language Models (LLMs) via Hugging Face to provide accurate, context-aware responses to user queries.