Posted on: 2026-04-09

I am currently leading a squad of four engineers, and we are working in an unconventional way: one week at a time. While short sprints have existed in the past, we are taking a slightly different twist.

Reasons

The reason I introduced this new work lifecycle was that we are developing a product with an AI-first mentality, using every AI resource available to expedite development. I identified a few bottlenecks in the conventional approach:

1. Too much manual work
2. Too many reviews
3. Focus on the output (code) more than on the end result (product)

The Process

The process is simple: I define the feature in a Google Doc alongside the product manager. Then, I tag the team. From there, we know what to do and who will do it. We have two Slack channels: one for developers only and one for specific features. The developer channel is for brief, candid, and egoless communication. The feature channel is for screenshots, screen recordings, sharing progress, and receiving feedback.

The process aims for a few philosophical aspects:

1. Speed: We rely on AI as much as possible with guardrails (for example, rules, linting, and unit tests).
2. Communication: We are a small squad that must remain aligned, necessitating quick communication on private Slack channels.
3. Unblock first: Anyone in the squad who spends more than 10 minutes on an issue (beginning to feel blocked) must raise a flag in the private Slack so we can resolve it.
4. United: We work in the same branch. We push and pull frequently and use AI to resolve code conflicts.

The Week

The week is structured so that Monday, Tuesday, and Wednesday are focused on AI coding. We code as fast as we can. On Thursday, we focus on better integration in preparation for the afternoon demo with stakeholders. We perform the demo and collect feedback. I re-prioritize with the product manager, and on Friday, we apply the required changes. Friday is a day with a different intention. I want the squad to breathe. Once the feedback is incorporated, we work on code quality. We ensure we do not have duplicates, we improve the rules, and so on. Finally, on Friday, we also explore and perform technical improvements, address technical debt, and try out new ideas.

Conclusion

The concept is a work in progress, and I will share the results in the future. So far, it balances productivity with clear periods of time for different speeds and purposes. Overall, using AI in this process and removing parts of a more traditional lifecycle significantly helps our velocity.