Multi-source knowledge ingestion
Pull references from **Notion**, Google Drive, Confluence, and Guru — then generate structured training from SME-approved documents.

Agentic GenAI Learning Management System
I architected and delivered CoTeach AI end-to-end — a multi-modal, agentic GenAI LMS that turns PDFs, emails, videos, and internal knowledge into publish-ready training in under 5 minutes (down from 2 weeks).
Built LangGraph and LangChain RAG architecture with Pinecone, LlamaIndex, and GPT-4 — ingesting PDF, DOCX, HTML, YouTube (Whisper), and 10+ sources (Google Drive, Notion, Guru, Confluence).
Delivered a FastAPI backend from scratch with 150+ API endpoints — RBAC, multi-tenant orgs, learning paths, assignments, cron jobs, and Slack/email notifications.
Shipped a responsive React frontend with a Notion-style block editor, AI chat assistant, and SME refinement workflows — reducing course creation from 2 weeks to under 5 minutes.
Deployed on AWS with CI/CD, Docker, and a multi-format ingestion pipeline designed for production scale.
Autoplays as you scroll · hover for sound
Scroll to preview · hover for sound
150+
API endpoints
2 wk → 5 min
Time saved
10+
Integrations
12
Languages
Screens from the product demo — ingestion, authoring, video modules, and global rollout.
Pull references from **Notion**, Google Drive, Confluence, and Guru — then generate structured training from SME-approved documents.

Block-based authoring with modules, quizzes, action steps, and one-click **Regenerate Training** — plus collaborator invites and publish flows.

Structured lesson paths with embedded quizzes, certification tests, and SME-friendly refinement — all editable in one workspace.

Loom-integrated video modules with instructor-led content inside each training — built for async and blended learning teams.

One-click training translation so global teams can publish once and localize for learners worldwide.

Interested in working together?
Get in Touch