Pydantic AI curriculum engine
Academic Architect
Academic Architect is a Pydantic AI-powered university curriculum engine for deans and professors. It generates pedagogically sound syllabi with Bloom's Taxonomy mapping, validated time-distribution logic, and professional grading rubrics.
Product walkthrough
Screens from a 10-day intensive Legal Tech workshop demo: configure the course, orchestrate modules with AI agents, refine day plans, and export LMS-ready JSON plus professor Markdown.
Course configuration
Deans and admins set subject, duration, session length, and choose local or cloud AI models.
Curriculum orchestrator
Phase 1 planner agent outputs module structure, learning objectives, and Bloom's Taxonomy mapping.
Module detailer — Module 1
Phase 2 detailer agent expands each module into day plans with timing breakdowns, materials, and rubrics.
Module detailer — Module 2
Professors review generated day plans and can regenerate individual days with feedback.
Final syllabus export
Dual-pane export delivers structured JSON for LMS integration and Markdown for print (Spanish locale).
English locale export
Locale switcher drives agent prompts and export headers; UI and Markdown follow the selected language.
Project highlights
- Two-phase Pydantic AI orchestration with output validators that enforce session timing programmatically.
- Local Ollama vs cloud model toggle for privacy-sensitive environments such as legal education.
- Dual export (LMS JSON + professor Markdown) with Spanish and English localization.