About

I run agents in production,
not just in demos.

I'm Brad — an AI engineer who came up through SEO and content strategy and taught myself to build serious systems. I don't just prototype: I run agents in production, unattended, every day, with the reliability engineering (self-healing, locking, rollback, evals) that keeps them trustworthy. That work now ships as real products — I founded AiVoicePhone, a 24/7 AI receptionist, and TreCal Media Group, an AI agency serving realtors and small business owners. I care about orchestration, tool/MCP integration, and making autonomous systems that people can actually depend on.

Approach

Production over prototype

Anything I build has to survive being left alone. Scheduling, retries, self-healing, and audit trails are part of the first version, not a later hardening pass.

Reliability engineering for agents

Locking, snapshots, rollback, verification gates, and evals — I treat flaky models the way a backend engineer treats flaky services.

Orchestration & tool use

The interesting work is in composition: routing between specialists, wiring MCP tools, and giving a whole system one dependable entry point.

Stack

Claude / Claude CodeMCP tool integrationsNotion as data layerScheduled cloud routinesNext.js / TypeScriptGSAP · Three.jsStripe · Gmail · Ahrefs APIsCustom CLI harnesses