I'm a B.Tech Computer Science (AI & ML) student at UIET with a passion for building production-grade ML systems. I specialize in deep learning, computer vision, NLP, and MLOps, with hands-on experience deploying real-world AI solutions.

• 🔭 Currently working on AI-powered cancer detection systems and large-scale ML pipelines
• 🔬 Research Intern at NIT Warangal | ML Intern at DRDO
• 🌱 Learning advanced Deep Learning, Neural Networks, Computer Vision & LLMs
• 🏆 Microsoft Learn Student Ambassador | Top-60 Healthcare AI Hackathon Finalist
• 📝 Published researcher with 2 ML research papers on degradation prediction
• 💬 Ask me about PyTorch, TensorFlow, LangChain, RAG Systems, MLOps, Computer Vision
• 📫 Reach me at deepanshusnpt@gmail.com
• ⚡ Fun fact: I build production ML systems that solve real-world problems!

Sep 2025 - Jan 2026

• 🌾 Developed end-to-end ML pipeline for crop output forecasting across 572 Indian districts
• 📊 Deployed LightGBM model achieving R² = 0.906, MAE = 17.65 kg/ha with 40× faster training
• 🛰️ Processed geospatial datasets (NASA POWER, MODIS, SoilGrids) with 35 engineered features
• 🎯 Established past crop outputs as dominant predictors through comprehensive feature importance analysis

June 2025 - Aug 2025 | Delhi, India

• 🔍 Built ML-based detection system for biochemical threat analysis using sensor data
• 🤖 Implemented ensemble models for anomaly detection with real-time inference capabilities
• 🚀 Collaborated with defense research teams on intelligent threat identification systems
• ✅ Achieved high-accuracy threat signature identification for national security applications

July 2025 | Tech Stack: LangChain, FAISS, Llama, Mistral, REST API

• 🎯 Engineered production-grade RAG pipeline mapping FIRs to 500+ IPC & 350+ BNS legal sections
• 📈 Achieved 92% classification accuracy with automated legal case classification workflow
• ⚡ Reduced manual review time by 60% and response times by 40% for law enforcement
• 🚀 Deployed REST API with LLM inference endpoints processing official government PDF knowledge base

March 2025 - June 2025 | Tech Stack: RoBERTa, Sentence-BERT, PyTorch, CNN, Hugging Face

• 🔬 Designed end-to-end production AI pipeline with fine-tuned transformers achieving 86% journal-scope accuracy
• 🖼️ Implemented CNN-based image classification for document quality assessment (DPI checks)
• 🤖 Orchestrated multi-modal PyTorch workflows for AI-content detection and plagiarism analysis
• ⚡ Serving 250 PDFs/day at < 2 min latency via optimized Hugging Face Inference Endpoints

Jan 2025 - Feb 2025 | Tech Stack: CNN, PyTorch, Computer Vision, GDPR/HIPAA

• 🎯 Built computer vision system for cancer detection achieving 92% classification accuracy
• 🔒 Implemented secure model serving infrastructure with GDPR & HIPAA-compliant deployment
• 🏆 Secured Top-60 position (Semi-Finalist) in Predictive AI in Healthcare with FHIR® International Hackathon
• 📉 Reduced data breach risk by 30% through enhanced security measures

⚠️ Note: Some advanced ML projects are in private repositories, undergoing research validation and production optimization! 🚀

1. 📄 Solar Panel Degradation Prediction using Machine Learning: A Comprehensive Approach

   • Status: Preprint | Read Paper | GitHub Repo
   • Machine learning models for predictive maintenance of solar energy systems

2. ⚡ Fuel Cell Degradation Prediction Using Machine Learning Models

   • Focus: Proton Exchange Membrane (PEM) Fuel Cell Dataset
   • Status: Preprint | Read Paper | GitHub Repo
   • Advanced ML techniques for fuel cell lifecycle prediction

• 🎖️ Microsoft Learn Student Ambassador - Recognized for technical leadership and community impact
• 🥈 Top-60 Finalist - Predictive AI in Healthcare with FHIR® International Hackathon (2025)
• 📖 Published Researcher - 2 ML research papers (Solar Panel & Fuel Cell Degradation Prediction)
• 🔬 Research Experience - NIT Warangal & DRDO

• 🏥 AI-Powered Cancer Detection Systems - Multi-modal diagnosis for lung, skin, oral, and cervical cancer
• 🧠 Advanced Deep Learning - Neural networks, image segmentation, and computer vision applications
• 🚀 Production ML Systems - Scalable model deployment and MLOps best practices

I'm always interested in collaborating on ML/AI projects, research opportunities, and innovative tech solutions. Feel free to reach out!