Tech Explained: The AI tool giving teachers time back in Simple Terms

Sarvesh Bhardwaj didn’t set out to disrupt American education. At first, he was just a parent watching his son struggle.

Even in a well-resourced school district, something felt off. His older child — a bright, motivated, college-bound middle schooler — was waiting weeks for feedback on assignments. Teachers were stretched thin. Essays piled up. Learning stalled in the delays between instruction and response.

“I started to see that there is a gap in terms of how much time it takes for teachers to be able to give feedback to students,” Bhardwaj says. “And I thought, if we could use artificial intelligence, we could make an impact.”

More than a concerned parent, Bhardwaj is a computer scientist trained to see patterns in complex systems and fix them. He earned his doctoral degree in electrical engineering from the Ira A. Fulton Schools of Engineering at Arizona State University in 2006, conducting research through the Center for Intelligent, Distributed, Embedded Applications and Systems, or IDEAS, Center. There, he worked at the early edges of artificial intelligence and machine learning.

When Bhardwaj recognized the widening gap between student effort and meaningful feedback, he didn’t just see a problem. He saw a system ripe for redesign. That realization became the seed for Magna Education, Bhardwaj’s fast-growing education technology startup that, in just a few years, has garnered national and international recognition, including being named one of District Administration’s Top EdTech Products in both 2025 and 2026 and earning Corporate LiveWire’s EdTech Platform of the Year, along with an invitation to the White House’s 2025 AI Education Task Force meeting.

The company’s rise reflects a broader shift in how AI is moving out of labs and into classrooms. But Magna’s approach is distinct. It doesn’t aim to replace teachers. It aims to give them time back.

Closing the feedback gap

Magna Education, which operates under the parent company Studygenie, focuses on one of the most stubborn problems in education: meaningful, timely feedback.

Traditional digital platforms lean heavily on multiple-choice questions because grading free-response work at scale is difficult. But real learning, especially in Advanced Placement courses, happens in essays, data analysis, math derivations and explanations.

“What we focus on is making sure that we create rubric-aligned assessments,” Bhardwaj says. “Then we grade it and give feedback to the student. ‘This is where you went wrong and this is how the College Board actually expects you to write it.’”

The system doesn’t stop with the student. Magna aggregates responses across an entire class, surfacing patterns teachers would normally need weeks to uncover.

“Something that would have taken two to three weeks, with feedback arriving a month later, can now happen in minutes,” Bhardwaj says. “That tight loop makes a difference.”

The results are striking. In one Louisiana school using Magna for AP courses, the percentage of students earning the highest possible score of 5 tripled — from 12% to 35% — in a single semester. A regression analysis comparing prior semesters confirmed the gains were statistically significant, increasing the odds of a top score by more than three times.

And the improvement wasn’t limited to top performers. Score distributions shifted across the board, lifting students at every level.

Before AI was a buzzword

Long before Magna, Bhardwaj was working on problems that now feel prescient. In the Fulton Schools, he conducted research under the supervision of Sarma Vrudhula, a professor of computer science and engineering in the School of Computing and Augmented Intelligence, in areas that bridged machine learning, probability and real-world systems.

“We were working on machine learning and AI at a time when those terms weren’t everywhere yet,” Bhardwaj says. “We focused on Bayesian networks, graph networks and using probability theory to predict properties of complex systems.”

That foundation shaped how he thinks and how Magna operates.

“I feel like having that background in research prevents you from throwing stuff out there without making sure it’s grounded in core principles,” he says. “We use a lot of data. That’s because I was trained at ASU to use data to inform decisions and then build something that is actually of value.”

That rigor also shows up in Magna’s emphasis on responsible AI. Rather than deploying generic models, the company works closely with subject matter experts and teachers to ensure grading aligns with established standards.

Lessons that last

Bhardwaj credits much of that mindset to his doctoral advisor. Vrudhula sees a direct throughline from Bhardwaj’s early work as a graduate student to the impact he is making today.

“Sarvesh always combined deep mathematical thinking with a strong sense of purpose,” Vrudhula says. “In taking rigorous research and using it to solve meaningful problems at scale; what he’s doing now is exactly what we hope doctoral training enables.”

For Bhardwaj, the mentorship went beyond technical skills.

“I think the most important thing I learned from Sarma was the love for math and making sure whatever you’re producing is really high quality,” he says. “Having worked with him helped me become the researcher and the individual who I am today.”

Building toward a bigger future

Magna is currently used in schools across California, Oregon, New Jersey, Louisiana and beyond, with plans to expand nationwide and internationally. While the platform began with AP courses, the long-term goal is broader.

“We want to expand to K-12,” Bhardwaj says. “We’ve built the infrastructure so we can take any state standards and create the curriculum and assessments very quickly.”

Five years out, Bhardwaj sees Magna not just as a tool, but as part of the backbone of AI-enabled education.

“We want to be considered pioneers in using AI responsibly for education,” he says, “not replacing teachers but empowering them to connect more deeply with students.”