.png&w=256&q=75)
1
1
China
1 year of experience
.png&w=828&q=75)
SmartTutor is an AI-powered personalized learning platform designed for K-12 students, addressing the widespread lack of affordable, one-on-one after-class tutoring support—especially in STEM subjects, where step-by-step guidance and curriculum-aligned practice are critical. Built to ease teacher workload and boost learning outcomes, the platform integrates intelligent homework assistance, adaptive question generation, textbook Q&A, and progress tracking for students and parents. At its core, SmartTutor uses a dual-loop multi-agent AI workflow: an investigation loop retrieves course materials, web resources, and citation-backed evidence, while a solving loop coordinates reasoning, tool use, code execution, and answer refinement. The system leverages RAG (Retrieval-Augmented Generation) to ensure responses align with official curricula, and Python code execution to verify STEM calculations accurately. The frontend is developed with Next.js and React, delivering a clean, student-friendly interface that supports mathematical notation and persistent learning sessions. The backend uses Python FastAPI with modular services for tutoring, question generation, user management, and analytics, alongside structured local storage for user data, activity logs, and learning reports. Key achievements include a fully functional core prototype, traceable multi-agent reasoning, hybrid retrieval tools, and age-appropriate explanation generation. Current progress stands at 55%, with ongoing work focused on UI refinement, expanded curriculum content, parent-facing progress dashboards, and final deployment preparation. By combining personalized support, curriculum fidelity, and transparent learning analytics, SmartTutor offers a scalable, accessible solution to transform K-12 self-directed learning.
10 May 2026