TENOR
AI-native LMS for Higher Education
- Year
- 2025
- Status
- in-progress
- Chapter
- Startups
- Role
- Founder & Lead Product Engineer
- Duration
- 2025–Present
- Disciplines
- Product Development, Business Strategy, UI/UX, AI Architecture
- Tags
- AI, EdTech, Startup, B2B, SaaS
Full story
The Problem
Universities are relying on infrastructure built in the early 2000s. Professors spend hours grading, navigating archaic UI patterns, and fighting with rich text editors that break tables. Students receive delayed, generic feedback and struggle to find course materials buried in deep folder structures. The friction of the tool gets in the way of the learning. We set out to build an LMS that feels like it belongs in the AI era—proactive, deeply personalized, and invisible when you don't need it.
The Architecture
We built the core engine on top of custom LLM pipelines, focusing heavily on RAG (Retrieval-Augmented Generation). The challenge was hallucination control. When a student asks a question about a midterm, the AI must strictly rely on the professor's syllabus, lecture transcripts, and uploaded PDFs. We implemented a vector embedding strategy that isolates course context into ephemeral sandboxes, meaning the AI knows exactly what context it's allowed to pull from at any given millisecond.
Current Status
Currently in closed beta with two private university pilot programs. The feedback loop is intense. We are learning that while AI grading suggestions save time, professors are highly skeptical of automated rubrics unless the reasoning is completely transparent. We are iterating heavily on explainable AI interfaces.


