I believe developers have a responsibility to use the best tools available.
We can build what founders imagine. We can turn ideas into working products. We can ship in weeks what used to take months.
I see this when working with early-stage founders. Their eyes light up when they see their product taking shape fast. "Wait, we can actually launch next month?"
With AI-assisted development, we can move faster without sacrificing quality.
I see that as a net positive—founders need to ship fast, and we can help them do that.
Yes, AI is changing how we code
But it's making us better developers, not replacing us.
AI handles the repetitive stuff—boilerplate, common patterns, basic features. That frees us up to focus on what actually matters: your business logic, user experience, and making sure everything works right.
We can explore more solutions faster. We can test different approaches without wasting days on implementation.
It also means more time for the stuff that matters—understanding what you're building, why it matters, and how to build it right.
The code still needs architecture. The features still need planning. The product still needs someone who gives a shit about whether it works.
AI just helps us get there faster.
Without AI:
12 weeks
With AI:
8 weeks
"Ok, but how exactly do you use AI?"
Every day. But not blindly.
a) We use AI tools like Cursor and Claude for development. They help us write cleaner code faster. But we're not just accepting whatever AI suggests—we review everything, refactor, and make sure it's maintainable.
b) For early-stage startups, AI lets us ship MVPs fast while keeping costs reasonable. You get a working product in weeks, not months. We focus on getting you to launch, not perfect code that never ships.
Our principles
- •AI is a tool, not a replacement for thinking
- •We use Cursor, Claude, GitHub Copilot daily
- •We still review every line of code
- •We still architect the system ourselves
- •We still make the hard technical decisions
- •AI handles boilerplate, we handle the interesting problems
- •We're always learning new tools (it's moving fast)
(this list will evolve as the tech evolves)
We mostly use Cursor, Claude, GitHub Copilot, and GPT-4 for development. We're always testing new tools. But AI is just one tool in our stack. It helps us work faster and smarter, but it doesn't replace experience, judgment, or caring about building something great.
Below you can see examples of projects we've built with AI-assisted development.
Examples of AI in Our Work
Project: FMCA Tax Filing Platform
Automated DOT tax filing system for transportation companies. Built a portal that generates compliance PDFs automatically—what used to take hours of manual work now happens in seconds.
Used AI to help scaffold the PDF generation logic and form validation. We handled the complex tax compliance rules and making sure everything meets DOT requirements.
Tech: React, Python, automated document generation
fmca-generator.com →Project: Katzkin VRS Pattern Management
Workflow management system for automotive leather production. Tracks patterns through 7 sequential stages with approval workflows and documentation at each step.
Used AI to generate the state machine logic and form structures. We designed the workflow architecture and business rules.
Tech: React, workflow automation, multi-stage approvals
katzkinvrs.com →Project: Healthcare Platform
React Native mobile app with real-time features. Used AI to generate boilerplate components and common patterns, we focused on HIPAA compliance and user experience.
Project: AI Integration
Added ChatGPT features to existing SaaS platform. Used Claude to help design prompt architecture and error handling. Shipped in 4 weeks.
How We Actually Work
Week 1: We talk about what you need. AI doesn't do this—we do.
Weeks 2-6: We build. AI helps us write code faster. We make sure it's right.
Week 7: We test everything. AI can't QA—humans do.
Week 8: We launch. You own all the code.
The result: Your product ships fast, works right, and you didn't waste 6 months waiting.