Content Generation
System (CGS)
A backend system for turning instructor requirements and reference material into course content. The production LMS path drafts PDF-grounded TipTap content; a separate course-agent workflow handles stateful multi-lesson generation. Developed collaboratively within Fusemachines Nepal's company-owned codebase during my AI Engineering Internship.
- Problem space
- Coordinate planning, lesson writing, review, revision, and rendering in a structured generation workflow.
- My role
- Contributed to backend and AI workflow engineering for the LMS authoring path and the separate course-agent path; did not independently create or own the full system.
- Technologies
- Python · FastAPI · LangGraph · LLM APIs · Pydantic · PostgreSQL · pgvector
- Instructor brief
- LMS frontend
- LMS backend
- CGS /author
- Async task
- pypdfium2 → LLM → TipTap
- Client polling