05 / Infolyze AI Suite · AI Engineer · Apr-Sep 2025 · Infolyze · United States
LinkedIn post ↗Four production
AI products.
One engineer.
Five months.
Designed and shipped four production AI products end-to-end as a core engineer at Infolyze.
Emmy
The inbox that learns who you are.
One OAuth handshake, then Emmy runs 24/7 · sorting the inbox, auto-creating labels that match the way the user actually thinks, drafting replies in their voice, and sending them. The sorting model improves with every email processed. Inbox management is not a task anymore; it is infrastructure.
Gmail OAuth · Continuous learning · 24/7 agent
Adsy
From brief to published in a single session.
Adsy generates branded ad creative and social posts · images and copy together - then schedules and publishes them across platforms without leaving the product. The hard part was not the generation; it was building the scheduling and publishing layer so that content production and distribution collapsed into one workflow.
AI image generation · Cross-platform scheduling · Publishing API
Dominic
Every registrar, one ranked list, live.
Domain hunting has always meant toggling between tabs and settling for second choice. Dominic queries all major registrars simultaneously, checks real-time availability, and returns a ranked list · scored by quality signals, not just alphabetical order. The right domain is at the top before the user has time to second-guess.
Registrar APIs · Real-time availability · Ranking engine
Revvy
Your Google Business Profile, never neglected.
Local reputation is slow to build and fast to lose. Revvy auto-replies to incoming Google reviews in a voice that matches the business, flags reviews that look fraudulent for manual review, and keeps the engagement signals that Google's ranking algorithm rewards. Location visibility compounds quietly over time.
Google Business Profile API · Review NLP · Fraud detection
Email infrastructure. Content generation and cross-platform publishing. Registrar API orchestration. Google Business Profile management. Four different APIs, four different problem shapes, four different user contexts · all inside five months, all in production, all for one client. Building across that surface is not generalism; it is what happens when one engineer is both the bottleneck and the leverage.
What connects Emmy, Adsy, Dominic, and Revvy is not the tech stack - each one touched different APIs, different data models, different user expectations. What connects them is a way of building: find the task that creates friction every day, build the agent that removes it, ship. The recurring pattern across four different domains is the evidence that this is not a method that only works once.