May 2026/Fullstack/professional
Teman Tuli
College lectures move insanely fast. If you're a Deaf student, falling behind on context happens in seconds. I built Teman Tuli because I was annoyed by how most captioning tools out there either have terrible real time accuracy or they just upload everything you say to the cloud without asking. I wanted to build a captioning app that actually respects privacy and sits right in your pocket.
To make sure the captions show up instantly without any lag, I relied entirely on Apple's native Speech framework and SwiftUI. It all runs right on the device. But the funny part was how I handled the backend. I literally made the Fastify and Prisma API as dumb as possible. It only exists to save transcripts when the user explicitly hits the save button. No automatic cloud syncing, no weird background tracking. Just total control.
Working on this really shifted my perspective on accessibility tech. It made me realize that helping marginalized users is not just about making text bigger or readable. It is about giving them full ownership of their data. When you are talking about sensitive stuff in a classroom, privacy is not a premium feature. It is an absolute must.