+ Developed a data retrieval platform combining NLP and SQL optimization to interpret user queries, generate dynamic SQL, and retrieve data accurately. Built with LangGraph framework for enhanced LLM capabilities. The system supports tabular data formats like XLS and CSV, includes chat history for context-aware responses, and features chart visualizations using Streamlit for interactive insights. Connected to cloud-hosted databases on Google Cloud Platform (GCP) using PostgreSQL or MySQL, and powered by Vertex AI for intelligent query handling.+ Result: The chatbot delivers good SQL query answers and provides interactive data visualizations using many type of charts/graphs, ensuring accurate and efficient data retrieval.+ TechStack: LangGraph, PostgreSQL, MySQL, Google Cloud Platform, Streamlit

+ Designed and developed an advanced chatbot system leveraging the LangChain framework and Django Rest Framework to create a robust, dynamic interaction platform. + Implemented long-term memory and conversation history features in RAG system, enabling the chatbot to provide more accurate and context-aware responses over extended interactions. + Integrated image processing capabilities using YOLOv8 and OCR libraries, allowing the chatbot to read, interpret, and understand content within PDF documents. + Deployed the solution using Docker for containerization and PostgreSQL for database management, ensuring a scalable and reliable deployment on the server.+ Langchain Framework: Proficient in using Langchain for building modular and scalable chatbot solutions.+ Langgraph: Used Langgraph for structuring and visualizing complex conversational flows.+ Django: Utilized Django as the backend framework for efficient application integration and deployment.+ Solve Document Layout Analysis problem with OCR and Yolov8 with contour.

+ Research, report about HINT: Hyper Instruction Tuning for Efficient Zero & Few- Shot Generalization.+ Conducted in-depth research on Optuna, a cutting-edge hyperparameter optimization framework, and integrated it into a stock prediction model to enhance performance and predictive accuracy.+ Research a Survey: Time-series data preprocessing: A survey and an empirical analysis. Report and preprocess on stock price data.+ Developed a strategy, conducted extensive research on Optuna and integrated it into a stock prediction model to enhance performance and predictive accuracy.+ Further improved the model using the Optimal Number of Clusters (ONC) algorithm based on price features to enhance the Sharpe ratio for a suitable approach.