kReative Labs
Back to kReative Labs

Live 2026

Ask me anything

A voice agent that answers questions about my work. Tap the microphone, speak, then pause. It answers out loud, and offers to take you to the longer version of whatever it just said.

A resume you can interrupt

Reading a portfolio is a linear act. You start at the top and work down through whatever order someone else decided. A conversation is not like that. You ask the thing you actually wanted to know, and if the answer is thin you push on it.

Voice removes even the typing. The interaction is closer to catching someone at a conference than to browsing a site, which is a better fit for the questions people actually have about a career.

It pairs naturally with the rest of this site. The agent-facing home page hands a visitor's model structured data to reason over. This does the same job for a person who would rather just ask.

Speech in, grounded answer out

Voice agent flow Speech is captured in the browser and transcribed. The question is matched against a knowledge base built from the career record. An answer is generated from what was retrieved, spoken back to the listener, and accompanied by a link through to the longer written version. Speak tap, talk, pause Transcribe speech to text Retrieve knowledge base built from the career record Answer grounded in what was retrieved Spoken out loud Full story link through
The link back to the written version is deliberate. A spoken answer is short by necessity, so the interface hands you somewhere to go when the short answer is not enough.

The flow above describes the interaction as it behaves from the outside. Worth extending with the actual speech and model providers, the retrieval approach, and how the knowledge base is built and refreshed.

At a glance

InteractionPush to talk. Speak, pause, and it answers.
AnswersGrounded in a knowledge base built from the career record.
OutputSpoken response plus a link through to the longer written version.
HostingRailway.
StageLive.