A student, a professor, and a problem worth solving

The problem

Every semester, the same pattern repeats across higher education. Students submit work and wait. When feedback finally arrives weeks later, it's often too generic to act on. The learning moment has passed.

Educators face an impossible choice: give every student the detailed, personalized feedback they deserve, or have a life outside of grading. With 150+ students per course, thoughtful feedback means 40-hour grading weeks. Something has to give.

Two perspectives

Laksh Agarwal was a computer science student at Simon Fraser University. He'd submit assignments and wait for weeks for feedback that might help him improve. When it came, it was usually a grade and a line: "Good effort, needs more analysis." By then, he'd moved on to new material.

Dr. Aishwarya Deep Shukla was on the other side. With a PhD from the University of Maryland and 7+ years teaching at SFU, he knew what good feedback looked like. He also knew the math didn't work. 180 students across three sections. 40+ hours per assignment if he gave each paper the attention it deserved. By paper 50, he wasn't the same grader he was at paper 1.

They ended up in the same research lab. Laksh was building machine learning systems. Aish was studying how people make decisions with information. For three years, they worked side by side on different projects, but kept circling back to the same question: could AI help close the feedback gap without taking educators out of the loop?

The experiment

They built a prototype. Rough, minimal, functional. Then tested it: comparing AI grades against human grades, refining rubric alignment, iterating based on what educators actually needed. The result was an AI system that could generate personalized feedback with one non-negotiable constraint: educators review and approve everything before students see it.

It was meant for their own use. A research tool. But word travels in academic circles. A mention at a conference. A recommendation to a colleague at another university. Educators started reaching out, asking if they could try it.

And they kept using it. Not because the interface was polished, but because it worked. Feedback that took weeks now took hours. Educators stayed in control. Students got responses while the material was still fresh.

The company

What started as a research experiment had become something people relied on. That's when Ednius became a company. Not because we wanted to build a startup, but because educators kept asking for it.