Browse categories
English
$
USD
Connect to your AI
I will build a complete ai development suite for your codebase with claude code
Savino Fiore
About this gig
Your team already codes with AI, but the agent forgets your conventions, closes tasks with broken tests and needs someone watching it. I build the complete suite that fixes this, so your developers spend their time on specs and reviews instead of typing code.
I'm Tech Lead at Finanz, a fintech app with 500,000 users. My team ships production code with Claude Code, and I published the workflow as open source (doe-kit and lexi). With the same team we went from one release every 30-60 days to one a week.
What I add to your codebase:
- CLAUDE.md with your architecture, conventions and commands
- - Custom skills and slash commands for your recurring tasks
- - Hooks that stop the agent from closing a task without green tests
- - Subagents specialized in code review
- - A unit test suite on your core modules: the contract the agent must respect
- - Documentation so your team can run it on its own
Stacks: Flutter and Dart, TypeScript (React, Next.js, Node.js, NestJS), Python.
Before ordering, send me your stack and the size of your codebase: I will tell you which package fits.
Get to know Savino Fiore
Savino Fiore
Flutter and AI Engineer, Tech Lead of a fintech app with 500K users
- FromItaly
- Member sinceDec 2020
- Avg. response time1 hour
Languages
Italian, English
Tech Lead at Finanz, a fintech app for financial education featured on Forbes and Repubblica: from 40,000 to 500,000 users, and from one release every 30-60 days to one a week. I build Flutter apps for iOS, Android and web, backends on Supabase and Node.js, AI features on the Anthropic and OpenAI APIs, and n8n automations with LLMs. I work test-first: specs and tests before code, with an agent-first workflow on Claude Code that I published as open source (doe-kit, lexi). You get code that is tested, documented and ready for production. I speak Italian and English.
My Portfolio
FAQ
Does it work with Cursor, Codex or Copilot?
The unit tests and the written conventions work with any coding agent. Hooks, skills and subagents are specific to Claude Code.
How do I receive the work?
As a pull request on your repository or as a patch file, with a short guide that explains every file I added and how your team uses it.
My codebase has no tests. Is that a problem?
No, it is the usual starting point. At Finanz I built the test suite from scratch on a codebase that had none. The Standard and Premium packages start from your most critical modules.
