Ahmed Kamal
I've been building software for years – backends, AI pipelines, game engines, and everything in between. I think in systems, I build from first principles, and I have a hard time leaving a problem alone once it gets interesting. The rest is below.
I've been building software for years – backends, APIs, full-stack apps, and eventually an entire AI-powered platform as the sole technical founder of a legal-tech startup. The product worked, we handled real clients, and somewhere along the way I learned that shipping great software is the easy part of building a company.
Before that, I freelanced on Upwork for four years – 41 projects, 4.9/5 rating. Go, Next.js, AWS, whatever the problem needed. Good work, good clients. But doing it enough times made something obvious: if I'm going to pour this much into building things, they should be things I actually want to exist.
So now I'm building AI solutions for LegalTech – starting with LegalSort, the document classifier I wished I had when I was a CTO. Plus a SaaS framework and whatever else I can't stop thinking about.
2025 - Present
Founder - LegalSort.ai
Building AI-powered document classification for legal teams. Taking everything I learned as a CTO and turning it into the tool I wished existed – this time with full ownership of the outcome, not just the code.
2022 - 2025
CTO - Legal-Tech Startup
Sole technical founder. Built an AI-powered due diligence platform from scratch – data extraction and AI-assisted review workflows. The product shipped and handled real client work. What it taught me: systems break, and the ones that survive are the ones you designed to fail gracefully. Also, that building software and operating a company are completely different skills – and now I know which one I'm better at.
2018 - 2022
Freelance Engineer - Upwork
41 projects, 4.9/5 rating. Go backends, Next.js frontends, AWS infrastructure. Useful training, but eventually it made something obvious – solving other people's problems is a good way to learn, but a poor way to build anything that matters to you.
LegalSort is an AI-powered document classifier built for legal teams – the tool I wished I had when I was running a legal-tech startup as CTO. Documents come in, AI figures out what they are, and lawyers spend less time sorting and more time doing actual legal work.
I picked this problem because I lived it. Working with large volumes of documents is genuinely hard – not hard in a "needs more engineering hours" way, but hard in a "the domain is complex and most tools don't respect that" way. That kind of problem is the only kind I find interesting enough to build for.
I start from foundations. If I can't rebuild something from scratch, I don't trust that I understand it – so I usually end up rebuilding it.
I'd rather be proven wrong and learn something than be comfortable and stagnate. The fastest way to understand anything is to try to break it.
That shows up in the notes – AI, LegalTech, software architecture, and deep dives into questions that don't have clean answers. You've been warned.