Agent-P emerged from a desire to reduce the administrative burden on educators, freeing up their time for students and research.
Key learnings during development include:
- Leveraging fetch.ai's agent framework for autonomous agents.
- Utilizing Groq's models for natural language understanding.
- Multiagent architectures for task management.
- Optimizing academic workflows using AI.
Agent-P is a multiagent system using fetch.ai’s framework and Groq's language models, developed in Python. The core agents include:
- Email Management Agent: Monitors .edu emails, summarizes communications, and sends daily digests for critical decisions.
- Lecture Preparation Assistant: Analyzes presentation materials and generates potential student questions and answers.
- Research Progress Monitor: Tracks research students, identifies roadblocks, and suggests support.
- Grading Assistant: Grades both MCQs and descriptive responses, providing statistical insights on student performance.
These agents work together, integrating with platforms like Canvas.
- Data Privacy
- Scalability
- Accuracy in Grading
- Agent Coordination
- fetch.ai
- Groq
- Python
- Toolhouse
- Vectara