Most people call me V. At UKG I work on the layer underneath the AI products, and right now that mostly means the MCP server behind the company's AI assistant. 15+ teams have plugged 60+ tools into it, and the assistant picks them up at runtime without anyone touching agent code. That's my favorite kind of work because it makes other teams' jobs easier, and they never have to know I exist.

Before the MCP server, my first stretch at UKG was making the AI platform observable. 45+ microservices now emit OpenTelemetry traces, so every call the assistant makes can be followed step by step, and incidents get resolved about 40% faster than they did. Along the way I rewrote the layer that scrubs PII out of those traces. It had been quietly eating our Google DLP budget, and the rewrite took 60% off that bill.

On the side I've built Gateway. Draw a cloud architecture on a canvas and it deploys, with real Terraform and real Helm underneath.

Before UKG I was closer to the hardware. At Universal Audio I wrote firmware for a 4-in/4-out audio mixer on a dual-core microcontroller, and before that I built the SMS service at TextBuddy so people without internet could reach an LLM by text. I spent a stretch on deep learning too, mostly rebuilding published models to understand how they worked.