Software engineer.
Product builder.

C taught me that the computer does exactly what you say.

Not what you meant. Exactly what you said. That gap — between intent and instruction — is where every bug lives. I've been thinking about it ever since.

500 problems. Most of them wrong on the first try.

Competitive programming on Codeforces, HackerRank, UVA. Nobody used the solutions. There was no user, no deployment. But the limits were real — time, memory, edge cases that break anything naive. I started to understand that a tight constraint doesn't block a good solution. It shapes one.

First time a teacher used software I wrote to mark attendance.

OnnoRokom Software Ltd. A school management system — students, attendance, grades. I remember realising the data in that database belonged to real children at a real school. A bug wasn't a failing test anymore. It was a teacher who couldn't do her job.

Someone's flight booking depended on an API I wrote.

A4Aero. I built the entire OTA platform — user, agent, admin — from scratch. Flights, hotels, payment gateways. The first production incident that woke me up at night changed how I think about uptime. It stopped being a technical metric.

Gave up a salary to find out if I could build something people need.

No co-founder. No funding. Started Ezdu and Better IELTS at the same time. When there's no one to escalate a decision to, you stop second-guessing and start deciding. Every architecture call, every product tradeoff, every deployment — mine. I found out I could.

Two products live. Harder than expected. More interesting than hoped.

Ezdu and Better IELTS in production, with real users. The technical problems got solved. The harder ones — what to build next, what to cut, how to stay focused when everything feels urgent — those took longer to figure out and are still being figured out.