A collection of research projects exploring different aspects of AI agents and their applications.
Investigates and compares different process types (Sequential vs Hierarchical) in CrewAI for conducting research and generating reports. This project demonstrates how different agent orchestration methods affect performance and output quality.
Implements and compares two different Retrieval-Augmented Generation approaches using CrewAI:
- PDFSearchTool for direct PDF document analysis
- CrewDoclingSource for knowledge-based document processing
The system combines document analysis with web search capabilities to perform comprehensive research and generate detailed reports. It includes performance comparisons between the two RAG approaches.
An AI-powered agent system designed to analyze code and generate detailed reports. The system can process source code, perform technical analysis, and produce comprehensive HTML reports with insights about the codebase.