Building and Operating Conviro: The Work Around the Model
Engineering lessons from building Conviro: tenant boundaries, retrieval quality, provider routing, tool permissions, and operating responsibility.
I build AI and commerce systems that hold up in production.
Hasan Gültekin, developer. Sole engineer behind Conviro and ConvergePilot, with 14 years in commerce platforms, technical SEO and automation.
What I work on
The same four areas keep coming back, on my own products and on client platforms.
- AI systems in production
- Retrieval, model routing, answer validation and tool authorization for support assistants that talk to real customers.
- Commerce platforms
- Magento, WooCommerce and WordPress stores: catalog imports, checkout, caching, and the debugging when a request gets slow.
- Technical SEO and measurement
- Core Web Vitals, structured data, and GA4 and Tag Manager tracking that survives a redesign.
- Automation and journeys
- Marketo, Mailchimp, n8n and custom journey builders designed to be safe to retry.
Currently building
Conviro
AI customer-support platform. Web chat, WhatsApp and Instagram in one inbox, with retrieval, answer validation and handoff to a person.
ConvergePilot
Turns an expert's working method into executable workflows a team can run and improve.
Working on something similar?
I take on a small number of platform, AI integration and performance projects. Tell me what you are running and where it hurts.