Payward's infrastructure team lead Kamo Asatryan credits OpenAI Codex with a six-month acceleration in their shipping timeline. His team builds the core inference infrastructure powering all Payward products, where speed is the primary engineering value. That is not a minor productivity gain. That is a structural shift in how fast a team can move.
The video makes the case through a concrete, named engineering org with a specific metric, not a vague testimonial. Asatryan describes how Codex integrates into the workflow of a team responsible for foundational systems, the kind of work where mistakes compound and velocity is hard to maintain. The argument is not that Codex writes code for them. It is that it removes enough friction to change what is possible in a given sprint.
What is worth reading in full is how a team working on inference infrastructure, not a greenfield app, applies an AI coding tool. Infrastructure work is notoriously resistant to automation shortcuts. If Codex moves the needle there, the implications for other engineering orgs are significant. The original video is the primary source.
[WATCH ON YOUTUBE →]