DEVANSH CHOUDHARY

Versatile AI developer skilled in crafting clean, maintainable code, with a knack for adapting to new technologies and strong problem-solving abilities for efficient project development. Capable of effectively self-managing during independent projects as well as leading a team.

RAG System with Llama3, LangChain, and ChromaDB

• Developed a Retrieval-Augmented Generation system for document-based question-answering.
• Utilized Llama3, LangChain, and ChromaDB to retrieve and generate accurate responses from a vector database.
• Enabled efficient information retrieval for real-time applications.

DoS/DDoS Detection for IoT Using Machine Learning

• Created a protocol-based deep intrusion detection system with 96.3% accuracy in classifying IoT traffic.
• Addressed data imbalance and overfitting using Python, TensorFlow, and Keras.
• Enhanced IoT security against DoS/DDoS threats with advanced techniques.

Lung Disease Classification Using CNNs

• Developed a CNN-based model for lung disease classification on X-ray images, achieving 91.45% accuracy.
• Leveraged pre-trained models like DenseNet201 and VGG16 with TensorFlow and Keras.
• Provided insights for clinical diagnostics through deep learning techniques.

OpenCV Motion Detection System

• Implemented motion detection using OpenCV with background subtraction and erosion techniques.
• Optimized for security and surveillance by minimizing noise in real-time analytics.
• Ensured robust performance for practical applications in dynamic environments.

• Microsoft Azure Fundamentals AI-900
• Generative AI Course Udemy
• Complete Python Bootcamp Udemy
• Power Platform Fundamentals from Microsoft
• Computer System Security IITK