From Late-Night Coding Frustration to Building an EdTech Startup

Let me describe a scene that every CS student or self-taught developer knows too well. It's 11pm. You're staring at a LeetCode problem. You read the solution three times. You understand it. You close the tab. Two days later, you can't recall a single thing about it.

That was me. I came into software engineering from a mechanical engineering background, having taught myself Python from scratch to break into the field. I loved building things. But technical interview prep felt like trying to memorize a dictionary: you could do it, but none of it stuck.

So I built LeetMotion to fix it. Not because I had a grand entrepreneurial vision, but because I was frustrated and couldn't find anything that actually worked.

The Problem Was Never the Problems

Most people assume that if you grind enough LeetCode problems, something clicks. And for some people, it does. But for a lot of us, especially those who didn't grow up treating algorithms as puzzles to play with, grinding alone doesn't build intuition. It builds anxiety.

The research on learning is pretty clear: passive reading and repetition without visual feedback or spaced reinforcement leads to shallow retention. Yet the dominant approach to algo prep is exactly that: read a solution, maybe watch a video, move on.

I wanted to understand why a sliding window works before memorizing the pattern. I wanted to see the recursion stack unwind, not just read about it. And I wanted the things I studied to actually show up in my memory a week later.

What LeetMotion Actually Does

LeetMotion started as a Chrome extension that injects a Visualize button directly into LeetCode problems. No switching tabs, no searching for a YouTube explanation. You hit the button and watch your algorithm run step by step, animated, right where you're solving the problem.

It now covers 150+ problems and growing. But the extension was just the beginning.

From there I built out the platform with two more layers. First, spaced repetition MCQs that surface the concepts you're weakest on right when your brain needs a refresher, not before and not too late. Second, an AI mock interviewer called Argos that runs you through realistic interview scenarios, gives real-time feedback on your explanation and approach, and doesn't just tell you whether your code is right.

The freemium model lets anyone get started with visualization for free. The paid tier unlocks the full practice system.

Building It as a Student with a Full Course Load

I won't pretend the timeline was clean. I built most of this while finishing my MS in Computer Science at Clark University, maintaining a 3.8 GPA, and completing a software engineering internship in fintech. The stack is Next.js, React, TypeScript, Prisma, and PostgreSQL, which I chose deliberately because I wanted the platform itself to be a portfolio piece, not just a side project.

The honest version of building a startup as a student is a lot of late nights, a lot of scope-cutting decisions, and a constant tension between shipping something useful and making it perfect. I chose useful, almost every time.

Boston made that easier than I expected. The student startup energy here is real. Winning a competition pitch for LeetMotion and reinvesting that prize money back into the platform was a turning point, not just financially, but in how I thought about the project. It stopped being a side project and started being a company.

What I Learned That Nobody Tells You

Building an edtech product is humbling in a specific way. Your users know exactly what good learning feels like, because they've been students their whole lives. They have strong opinions and zero patience for friction.

A few things I'd tell anyone building in this space:

1. Your first version will solve a problem you have, not the problem your users have. Get it in front of people fast and expect to be surprised by what they actually need.

2. Distribution is harder than the product. I spent months building features before I seriously invested in how people would find LeetMotion. That was the wrong order.

3. Being the user is your biggest advantage early on, and your biggest blind spot later. The moment you stop struggling with algo prep the way your users do, you need external feedback loops.

Where It Goes from Here

LeetMotion has active users and a growing community. The next phase is expanding the visualization library, improving Argos, and investing more seriously in the content and distribution side of the business.

If you're a developer preparing for interviews, a bootcamp or CS program looking for a tool that actually helps students retain what they learn, or someone who just wants to finally understand why binary search works before memorizing it, come try it at leetmotion.com.

And if you're a founder in the Boston ecosystem building something similarly niche, something that started because you were frustrated: keep going. The problems worth solving usually start exactly that way.

About the author: Sumanth Aitham is a software engineer, MS CS candidate at Clark University, and founder of LeetMotion. He has 3+ years of experience across distributed systems, fintech, and full-stack engineering. He built LeetMotion after teaching himself to code and realizing that grinding LeetCode problems was the worst way to actually learn them