A software engineer in Lahore with five years of experience, all of it with Australian and US startups. Six months ago the Australian startup he worked for shut its Lahore office — the work, they said, was increasingly done by AI, and they would rather train the team in Australia. Everyone was laid off.
Since then: more than two thousand applications and twelve or thirteen interviews, none of which converted. He kept building in the meantime — an open-source project with a thousand GitHub stars, an app he eventually took down because he could not keep paying the API costs.
A friend referred him to a company in Lahore. The technical interview went perfectly; the coding test was easy. The final round was with the CEO, the CTO, and a manager. It was going well until the manager asked how he uses AI. He described his workflow — planning prompts, his own tooling, a video-editing tool he had built for himself. The manager asked whether a person with good analytical thinking could be a good engineer while using AI. He said yes. He could see the manager did not agree, and wanted something he could not identify. A few more AI questions, and the round was over.
Looking back, he thought he should have separated the two — that a programmer still has to debug, plan, and review a design before implementing anything.