AI Teaching Deck Generator for EdTech: Extramarks Case Study
Instant, curriculum-aligned teaching decks created from minimal teacher input. Integrated into the Extramarks platform for teachers and academic content teams.
Extramarks needed a way to generate classroom-ready teaching decks without forcing teachers and academic teams to build every lesson manually. MythyaVerse created an AI deck generation pipeline grounded in Extramarks' existing videos, slides, and curriculum-aligned materials. A RAG layer retrieves verified content, while a Gemini-based generation layer structures lessons into usable teaching decks with consistent pedagogy. The system integrates into the Extramarks platform, supports contextual customization, and includes an academic feedback loop so content teams can keep improving quality as usage patterns emerge.
Business Impact
The generator reduced blank-page lesson preparation and helped teachers start from a structured, curriculum-aligned deck. Extramarks gained a reusable content automation layer that supports consistency across academic teams.
Overview
Extramarks needed a fast, accurate, and consistent way to generate full teaching decks for classrooms. Teachers often spent hours creating slides manually, and content teams faced challenges maintaining consistency across subjects and grade levels. The goal was to build an intelligent system that could create complete, curriculum-aligned decks using a combination of teacher prompts and Extramarks' extensive content repository.
Problem
Teachers and content teams were spending significant time preparing slides while trying to maintain consistency across subjects, grades, and curriculum requirements.
Constraints
- Keep generated decks grounded in the existing Extramarks content repository.
- Respect curriculum alignment and pedagogy standards across subjects and grades.
- Support teacher-specific prompts without losing content consistency.
- Integrate into the existing platform rather than creating a separate workflow.
Solution Architecture
- Videos, slides, and learning materials are indexed in vector databases for retrieval.
- A RAG layer selects relevant curriculum-grounded source material for each deck.
- A Gemini prompt system organizes retrieved content into teaching-friendly slide flows.
- Platform integration and academic review feedback support continuous improvement.
The Challenge
Creating classroom-ready slides is time consuming. Teachers often build decks manually, resulting in inconsistent structure, varying quality, and significant preparation time. Content teams also needed a way to maintain uniform pedagogy across subjects and grades while supporting personalization when required. Key challenges included:
- Manual slide creation taking several hours per lesson
- Difficulty maintaining consistency across subjects
- Large volumes of content that had to be reused efficiently
- Need for topic-aligned, curriculum-compliant slides
- Requirement for contextual customization based on specific teaching needs
- Integration into an existing platform used at scale
Objectives
- Reduce preparation time for teachers and content teams
- Generate complete, ready-to-use teaching decks from minimal inputs
- Maintain consistent pedagogy across subjects and grade levels
- Use Extramarks' existing content library effectively
- Ensure curriculum alignment with high accuracy
- Support customization for specific contexts or teaching scenarios
Our Approach
MythyaVerse engineered a full pipeline that transforms teacher prompts into structured, curriculum-aligned teaching decks. The system leverages Extramarks' existing videos, slides, and content repositories, stored and indexed in vector databases. A RAG layer ensures each generated slide is grounded in verified content. A carefully crafted, domain-specific prompt was designed for Google Gemini models, ensuring clarity, pedagogy-friendly flow, and consistent structure.
Solution Delivered
- Teaching Deck Generation Engine -Full decks generated automatically using a mix of retrieval and generation.
- Curriculum Mapping and RAG -Content grounded in curriculum-aligned videos, slides, and materials stored in vector databases.
- Custom Prompting for Educational Flow -A specialized prompt designed for teaching logic, learner engagement, and clarity.
- Content Personalization -Support for teacher-specific context, examples, and lesson goals.
- Platform Integration -Seamless integration into the existing Extramarks platform with no disruption to workflows.
- Academic Review Loop -Continuous feedback from Extramarks academic teams to improve quality and accuracy.
Technologies Used
Google Gemini
FastAPI
MilvusRelated MythyaVerse Services
RAG Development
Build production RAG systems with MythyaVerse: enterprise retrieval pipelines, hybrid search, multilingual assistants, secure deployment, evaluation, and observability.
AI Automation
Automate business workflows with enterprise AI automation from MythyaVerse: support automation, content pipelines, document intelligence, dashboards, and integrations.
Build an AI education workflow
Talk to MythyaVerse about a production-grade AI, automation, or XR system with the right constraints designed from day one.
Explore More Projects
Extramarks -Classroom Activity Generator
Curriculum aligned, hands-on classroom activities generated instantly from teacher inputs. Integrated into the Extramarks platform for teachers.
ZebPay -Support Bot & Automation Hooks
A 24x7 intelligent support system for crypto users with automated resolution and human escalation. Deployed on web and mobile for all ZebPay users.
boAt -AI Lyric Generation Campaign
A Women's Day campaign that turned personalization into respectful, shareable storytelling. Users created custom song lyrics for women in their lives.