Led the development of a multi-agent bot leveraging CrewAI and LangGraph. Integrated a Retrieval-Augmented Generation (RAG) pipeline, reducing bot hallucinations by 95%. Developed features for retrieving and analyzing data from complex systems to provide informed financial insights. Implemented functionality for conducting blockchain transactions. Additionally, integrated inference capabilities from multiple LLMs, including GPT-4, LLaMA, Claude, Mistral, and PhiDeveloped an advanced… Show more Led the development of a multi-agent bot leveraging CrewAI and LangGraph. Integrated a Retrieval-Augmented Generation (RAG) pipeline, reducing bot hallucinations by 95%. Developed features for retrieving and analyzing data from complex systems to provide informed financial insights. Implemented functionality for conducting blockchain transactions. Additionally, integrated inference capabilities from multiple LLMs, including GPT-4, LLaMA, Claude, Mistral, and PhiDeveloped an advanced time series forecasting tool with over 90% accuracy, integrating statistical and deep learning models. Implemented a robust stacking ensemble architecture. Created a user-friendly interface for easy data upload, pre-processing, and forecast generation, reducing turnaround time by 50%. Show less

Assisted in the development and updating of course materials, ensuring alignment with the latest advancements in AI. Additionally, I managed grading responsibilities and conducted office hours to support student learning for the following courses:Deep LearningData Mining

Led the development of a Hybrid RAG architecture leveraging Neo4j knowledge graphs for context retrieval, creating an enhanced pipeline that utilizes relationships between documents and entities to improve retrieval efficiency and observability. This enabled stakeholders to efficiently access information from their documents with proper sourcing. Developed a chatbot designed for policymakers to conduct research, significantly reducing LLM hallucinations and enhancing the reliability and… Show more Led the development of a Hybrid RAG architecture leveraging Neo4j knowledge graphs for context retrieval, creating an enhanced pipeline that utilizes relationships between documents and entities to improve retrieval efficiency and observability. This enabled stakeholders to efficiently access information from their documents with proper sourcing. Developed a chatbot designed for policymakers to conduct research, significantly reducing LLM hallucinations and enhancing the reliability and accuracy of the chatbot. Show less

Integrated GPT-4 and Multi-Platform Chat Systems: Spearheaded the integration of OpenAI's GPT-4 with advanced prompt engineering and function calling, creating a multi-functional AI agent. Simultaneously integrated chat platforms such as Telegram, Discord, WhatsApp, and Facebook Messenger, resulting in a 150% increase in user productivity and a 400% boost in user engagement.Chatbot Development and Personalization: Acquired essential competencies in chatbot development, including… Show more Integrated GPT-4 and Multi-Platform Chat Systems: Spearheaded the integration of OpenAI's GPT-4 with advanced prompt engineering and function calling, creating a multi-functional AI agent. Simultaneously integrated chat platforms such as Telegram, Discord, WhatsApp, and Facebook Messenger, resulting in a 150% increase in user productivity and a 400% boost in user engagement.Chatbot Development and Personalization: Acquired essential competencies in chatbot development, including designing conversation flows, implementing user data storage, and leveraging personalized user experiences. Applied these skills to enhance the effectiveness and user engagement of chatbot applications.Database and Data Management: Orchestrated the management of user data by leveraging MongoDB and Firebase, ensuring efficient data handling and security across multiple platforms. Show less

Developed and implemented an efficient data processing pipeline leveraging Apache Spark, enhancing scalability and reducing processing time by over 30%. This enabled real-time analytics on data.Developed time series models with exogenous variables, improved forecast accuracy by 15%. This project enabled better decision-making by providing reliable, data-driven insights.Developed a decision tree-based machine learning model to predict customer churn by analyzing usage patterns… Show more Developed and implemented an efficient data processing pipeline leveraging Apache Spark, enhancing scalability and reducing processing time by over 30%. This enabled real-time analytics on data.Developed time series models with exogenous variables, improved forecast accuracy by 15%. This project enabled better decision-making by providing reliable, data-driven insights.Developed a decision tree-based machine learning model to predict customer churn by analyzing usage patterns and engagement data. Improved churn prediction accuracy by 20%, enabling targeted retention strategies and increasing customer lifetime value. Show less