I build new web products by directing AI agents through an engineering system of my own.
Bring the product problem, the people it must serve, and the constraints that matter. I define the architecture and acceptance boundaries; agents do much of the direct implementation work; I take responsibility for the software we accept.
TeqFW, modular architecture, JavaScript across browser and server, dependency injection, Node.js and Web Platform capabilities form an engineering system developed through real work—not a technology menu.
I change that system when working software reveals a better way to make the next change controlled and useful.
AI agents are my primary development interface
I no longer write most implementation code by hand. I direct agents through analysis, implementation, testing, and documentation while I set intent, constraints, decomposition, verification, and acceptance.
Build something new, or use something already built.
The same engineering practice supports both paths. They start with different needs.
Build a product with me
Bring a new product problem, its users, capabilities, and real constraints. I take responsibility for the architecture and engineering approach.
PDE is a practical product within the Alarisa direction; smaller systems also show the same thinking across publishing, controlled digital resources, and infrastructure.
Current Work now connects status and Journal evidence
Repurposed the retained Projects page to separate active systems from historical work and connect dated Events to current status without changing that status through chronology.
New article: Agent Skills inside TeqFW packages
How to ship instructions for an AI coding agent with a package and keep them aligned with the installed code version.
My Russian-language book about the browser is now published
A new book about the browser as a personal operating environment for the internet is available in Russian.
Wired Geese is my practice, not an artificial agency.
You work directly with the engineer who makes and accepts consequential decisions. The name describes a growing set of AI-agent capabilities working under that human responsibility.
Which path fits your need?
Bring a new product problem to build, or ask about deploying PDE or Telegram Desk for your own account and compatible agents.