Team
Real engineers, named, locatable.
Two people. No subcontractors, no offshore delivery teams. Every engagement is staffed by one of us.
Our story
Built by engineers who got tired of watching AI demos die in staging.
47hq started with a frustration. Usman and Ibrahim — brothers, both engineers — kept getting pulled into the same conversation: a company had paid a six-figure invoice for an AI prototype, the demo wowed the board, and then it never made it to production. Or it shipped, broke quietly, and nobody could explain why.
They had seen it from the inside. Usman shipping AI surfaces at B2B SaaS scale-ups, owning the eval frameworks and reliability discipline most agencies skip. Ibrahim running scoping conversations and watching well-intentioned engagements lose the plot the moment money changed hands. They were convinced there was a better way to do this work — one that started with a real diagnostic instead of a sales pitch.
We named the company after the time of day everyone's honest — 4:47 in the morning, when the on-call page goes off and you find out whether what you shipped was real. That's the bar. Production AI. Named owners. Fixed scope. Metrics you'd publish on the wall.
Two people. No subcontractors. No offshore handoff. Every engagement is scoped, built, and shipped by the same people you meet on the first call.
Usman leads the production-engineering practice. Before 47hq he shipped AI surfaces at two B2B SaaS scale-ups, owning the eval framework and reliability discipline behind them.
Focus at 47hq: reliability architecture, eval methodology, and the post-handoff operational discipline that keeps shipped systems trustworthy.
Ibrahim leads engagement scoping and the consultation product. He runs first contact on every engagement and writes the 1-page diagnostic that anchors the consultation.
Focus at 47hq: keeping scope honest, making sure we say no to bad-fit work, and translating reliability findings into language CTOs can take to the board.
What we ship
Reliability discipline, not demos.
Every engagement leaves behind the operational scaffolding your on-call rotation will actually use. Eval gates that block bad deploys. Runbooks tied to alerts. Rollback paths that fire.
Next step
Want to talk?
Book a paid consultation. One of us will run the session.
Refundable if we're not a fitWritten diagnostic in 48 hoursSession run by a founder, not a sales rep