Irtaza is the rare engineer who treats the boring half of a system (the runbooks, the migrations, the on-call rotation) as the actual product. We hired him for a six-week rewrite and kept him on retainer.
We had a working LLM prototype and a vague sense it shouldn't go to production yet. Three weeks later we had an eval harness, a gateway, a cost graph, and a very clear go decision. Worth every dollar.
He took our 11-year-old ETL graveyard and made it boring. The thing I appreciated most was that he refused to be precious about the rewrite: small steps, every one of them shippable.
Operates like a one-person product team. Writes well, draws well, ships on time, and is unusually pleasant during the part of a project where everyone usually starts blaming each other.