What I build
- SaaS MVPs with sign-in, dashboards, a database and the one core workflow your users pay for.
- AI features on top of LLM APIs such as Claude and Google Gemini: writing and rewriting, classification, summaries and content pipelines.
- No-code apps when they are the faster path to a working product.
- Browser extensions and integrations that connect your app to the tools people already use.
How I build: AI-assisted development
I build with AI-assisted development, sometimes called vibe coding, using Claude Code and Cursor. I describe each feature, let the AI write a first version and iterate on it. This is how one person can ship a full product, with a frontend, an API, a database and integrations, in the time a small team would need.
Speed is not the whole job. Every build is tested on real data and handed over with documentation, and security pieces such as password hashing, session tokens and rate limiting are written in explicitly, not left to chance.
Examples
SignPro is a legally binding e-signature platform built with React, Node.js and PostgreSQL. Senders upload PDF or Word documents and request signatures by email or SMS; every signing is sealed with an evidence hash. SociPRO is an autonomous social media agent built with Node.js, Google Gemini, Supabase and FFmpeg. Both have live demos and public source code.
I also built EurekaEDU, a scholarship platform whose listings are collected by an n8n and Google Gemini pipeline, and LinkedIn Bioster, an AI app that rewrites LinkedIn headlines and bios.
Who it's for
Founders who want to test an idea with a real product instead of slides, and teams that need an internal tool built around how they already work.
