[AirClass Dev Log 1] Why I Started AirClass: Motivation and First Implementation Plan
Why I began thinking about AirClass from the limits of tablet-based classes and existing learning platforms, and what I wanted to implement first.
Why I began thinking about AirClass from the limits of tablet-based classes and existing learning platforms, and what I wanted to implement first.
Why I chose FastAPI and Svelte for the first AirClass prototype, and why I first considered a teacher-laptop local-network deployment model.
Why I first tried to build AirClassDisplay so students could see and revisit the lesson screen more smoothly, and why it did not become a practical classroom tool.
How question-quality data, Bloom, and Dewey shaped MAICE's problem-definition strategy and multi-agent architecture.
How MAICE split question classification, clarification, answer generation, and learning observation into a Redis Streams-based multi-agent flow.
How MAICE lowered the question-entry barrier with formula input, OCR, and streaming-safe LaTeX rendering.