E-commerce Platform

Swelo adopts a Claude Code-centered development lifecycle

After Ravn introduced a Claude Code-centered SDLC and embedded a forward-deployed AI Engineer to train Swelo's engineering team on it, the share of Swelo commits referencing a tracked ticket, touching a test file, and passing a structured review gate went from near zero to a over 60% within weeks, a git-verified shift that tracks directly to the new lifecycle.

Interested in AI Transformation?

The Challenge

Swelo's engineering team had been shipping to production without a defined process around it. Work rarely started from a tracked ticket, tests were seldom written alongside features, and there was no structured review step between a Claude-assisted change and production. The team lacked a lifecycle that made that code traceable, tested, and reviewed.

The Solution

Ravn designed and introduced a new SDLC built around Claude Code: requests are shaped into tracked tickets before work begins, tests are written and run alongside each feature, and every change passes a structured review verdict and a "gate green" verification check before shipping. Rather than handing over a process document, Ravn placed a forward-deployed AI Engineer inside the Swelo engagement to train the team on the lifecycle in their own codebase, through working sessions and a dedicated "Swelo AI Workflow Walkthrough" session. That engineer continues to work alongside Swelo's team as the SDLC matures.

The Outcome

- Commits referencing a tracked ticket (SWE-nnn): 3% (all history before Sept) → 70% (Sept 17–22). - Commits touching a test/spec file: 13% (all-time before) → 60% (after); by month, a clean ramp from 0% in May to 48% by September - Commits carrying a structured "Review: PASS/FAIL" verdict: never appeared before Sept → 60% of commits Sept - Commits passing a "gate green" check: never appeared before Sept → present in his commits from Sept

Tech & Cloud

Claude CodeClaude Code

MORE WORK