This paper investigates the use of large language models for automatically generating educational content, with a focus on specifying precise, pedagogy-aligned requirements for AI-assisted authoring. We introduce EMGen, a web-based system that combines prompt design patterns with an iterative generate-review-revise process, grounded in a structured curriculum model and specialized educator-AI interfaces. These interfaces translate instructor intent into actionable constraints (e.g., content type, learning methodology, and question validation criteria) and, when paired with instructor-provided resources via retrieval-augmented generation, help align outputs with curricular goals. Through initial testing, we illustrate how EMGen supports the creation and refinement of materials; formal instructor evaluations of UI/UX and content quality are forthcoming. Our contribution is a practical workflow and toolchain that reduces ad hoc prompt engineering, keeps educators in control of pedagogical decisions, and shows promise for improving the quality and relevance of AI-generated educational materials.