Building seoaiagent.ai

I started by doing SEO service work myself: talking to business owners, delivering local SEO tasks, and learning which parts of the workflow were repeated enough to become software.

The first product direction was a self-serve dashboard for business owners, with progress tracking, review workflows, posts, and AI replies. That taught me that most owners wanted outcomes, not another tool to manage.

That pushed the company toward the current direction: an AI teammate for local SEO agencies, built around agency conversations, founder outreach, and the real work needed to serve local businesses.

Liart.io

A multiplayer browser game that became the most visited stall at LOCUS and reached more than 1,000 players.

Sceneverse fine-tuning pipeline

A complete loop from raw 3D scenes to text data, fine-tuning, experiment tracking, parsed predictions, and 3D output.

Patan Durbar Square reconstruction

Cultural heritage 3D reconstruction work later published in Digital Applications in Archaeology and Cultural Heritage.

Virtual shoe try-on

Turning short shoe videos into 3D reconstructions that could be cleaned, converted, and placed on a user's feet in mobile AR.