Portfolio Projects

Featured

AI Hotel Booking Assistant

Developed a Llama 2-based AI assistant using Docker, and HuggingFace, demonstrating practical applications of NLP in the hospitality sector.

NLP Docker HuggingFace
View Project
Featured

Pong AI with Deep Q-Learning

Implemented neural networks for game AI using PyTorch, showcasing advanced reinforcement learning techniques. This project demonstrates deep understanding of machine learning algorithms and their practical implementation in interactive environments.

PyTorch Q-Learning Game AI Deep RL
GitHub Code Publication
Featured

Finance RAG Implementation

Built a Retrieval Augmented Generation system for financial due diligence using GEMMA and Llama models.

RAG LLM MongoDB
View Project

Chatbot App with AWS Endpoint

Developed an interactive AI-powered chatbot platform using AWS Lambda and API Gateway. This web application features real-time responses, message history for context-aware interactions, and a responsive design that adapts to different devices.

  • Built with Python, Streamlit for the frontend, AWS services for backend
  • Implemented serverless architecture using AWS Lambda for scalability
  • Used LangChain framework for advanced LLM capabilities
AWS Python Streamlit LangChain Boto3
View Project

Gemini CLI + MCP Documentation Generator

Developed a versatile GitHub Actions workflow that automatically generates documentation using Google's Gemini CLI with Model Context Protocol (MCP) integration. This project demonstrates advanced CI/CD automation and AI-powered documentation generation.

  • Cross-platform support for Windows and Linux workflows
  • GitHub MCP server integration for enhanced context and automation
  • Configurable paths and flexible triggers for documentation generation
  • Comprehensive error handling and validation systems
GitHub Actions Gemini CLI MCP CI/CD JavaScript Automation
View Project

LlamaEval: Hackathon LLM Evaluation Dashboard

Developed a rapid prototype during a hackathon for evaluating and comparing Llama models using the TogetherAI API. This project demonstrates quick prototyping skills and practical LLM evaluation capabilities.

  • Built a user-friendly dashboard for model selection and benchmark tasks
  • Implemented performance metrics including accuracy and BLEU scores
  • Deployed on Azure Container Apps for cloud accessibility
  • Collaborated with team members in a competitive hackathon environment
Python Streamlit TogetherAI API Azure Docker LLM Evaluation
GitHub Code Live Demo