AI-enabled Intelligent Assistant to Improve Reading and Comprehension Skills in English Language
- Lack of dynamically adaptable excersises.
- Lack tools to provide assessments for essay-type questions that typically require human evaluation.
- Provide a platform to improve advanced comprehension skills such as Sequencing, Summarizing, Self-questioning
- Implement Agentic RAG workflow to generate dynamic paragraphs and exercises
- Implement Agentic RAG workflow to accurately evaluate answers and provide feedback without human involvement
- Develop the most suitable prompts to generate dynamic exercises, evaluate answers and provide feedback.
- Frontend: ReactJS
- Backend: Python, Flask, Langchain
- AI:
- LLM Model - Phi-3
- Inference Engine: Ollama
- Architecture & Patterns: ReAct, Function Calling
- Prompt Engineering Techniques - Zero-Shot Prompting, Few-Shot Prompting, Guided Prompting, Role-based Prompting
- RAG - Corrective RAG
- Database: ChromaDB
- Other: Google Drive
