The Handoff Is Dead. The Future of Design and Development.
How I shipped 45 PRs across iOS, Android, and server as a designer
Hey everyone,
It’s been a minute since my last post — Cash App has been moving at lightspeed. But this one felt important enough to make a post.
The launch
Recently, we shipped Cash App managed accounts for kids — a new way for families to manage money together. It rolled out to 100% of customers across iOS, Android, and server. Dive in to learn more.
That launch was a defining moment. Not because of what we shipped — but because of how we built it. Something fundamental had changed in the process, and it’s been invigorating.
The shift
The industry has been moving toward tighter design-to-engineering loops. But inside complex, multi-platform systems like Cash App, shipping production code as a product designer still felt out of reach.
I’d been watching that start to change internally. But this latest project is where I experienced it firsthand — and it wasn’t incremental. It was a fundamental shift.
Why this was different
Across this project, I personally merged 25 pull requests across three codebases — Android (14 PRs), iOS (8 PRs), and Server (3 PRs). As a designer.
And zooming out further — over the last two months, I personally authored ~45 PRs merged across 5 repos. As a product designer:
As the design lead, I’m in the code alongside the engineers. Not theoretically. Not “reviewing.” Actually in it — shaping the experience all the way to production.
Here’s the workflow that made it possible:
The build process: Where everything changed
Figma is still the source of truth. Product direction, design decisions, interaction specs — that’s still where it starts. What changed is everything that happens after.
1. Figma → Swift prototypes
I build functional Swift prototypes that bring my Figma designs to life — real interactions, real animations, real feel. These aren't throwaway demos. They become design review artifacts and the reference implementation that production code is built from — either by me or by engineers on the team.
2. Prototypes → Production code
Using Claude Code or Codex, paired with our internal tool Builderbot — which deeply understands our existing codebase — I transpose my prototypes and Figma specs into production code across platforms. Engineers review and merge.
3. The inflection point: Design leading the last mile
This is where the traditional dynamic flips. Instead of filing bugs and waiting for engineering to prioritize them, I lead the charge — logging issues, kicking off fixes, and drive them to completion.
The process: visual feedback and bugs drop into a dedicated Slack channel. Builderbot — which intimately knows our codebase — picks them up and posts rapid fixes, often within minutes.
I pull the code down, test on a real device, reference my Figma file via MCP, and approve or adjust until it’s right. Every PR still gets a human review.
This is the inflection point. The designer isn’t on the sidelines filing tickets anymore — the designer is in the driver’s seat, leading the last mile to ship.
The last mile. The difference between “shipped” and “crafted.”
What this changed — For me and the team
The speed was remarkable. But the real shift was what it did to the dynamic. When I started filing bugs AND fixing them through this workflow, the whole team’s velocity ramped to a new level. Engineers focused on deep architectural work with Design, PM focused on design-layer issues directly where fit.
The ownership felt different. The iteration cycle is so fast I can see changes live almost immediately. The thing I imagined is the thing that ships. That kept me in flow state for days.
For designers: this isn’t AI replacing us. It’s AI removing the walls between our vision and the final product. The role isn’t shrinking — it’s expanding.
Looking ahead
This process isn’t a one-off. We’re not only using it to ship new products and features — we’re using it to refine our app as a whole. Bug fixes, polish passes, interaction improvements — the same workflow that launched a feature is now continuously raising the bar across the entire product.
The direction is clear. When you pair human taste with AI velocity, the result doesn’t just ship — it keeps getting better.
This launch turned up the flame dial another notch and i’m hungry for more. 🔥🚀
Special thanks to Cam Worboys, Gustavo Balestraci, Alex Ourbari, and many more.
If you’re new here, I write about the future of design, AI, and building at the cutting edge. Subscribe so you don’t miss what’s next.