Projects
What I'm building, why, and how.
kkya labsLive
WhatA platform for merchants to understand how AI agents actually shop.
WhenApr 2026 →
HowControlled shopping trips with AI agents (2,000+ so far), translated into merchant findings.
WhyCommerce was built around humans. When the shopper is an agent, the old assumptions break.
- 2,000+ controlled agent trips; two production partner integrations. → kyalabs.io
- Built on our own agentic harness — most of what kya ships is built by agents.
- The thesis, long-form: Building in 2026.
- What it’s teaching me: agent capability changes month to month, and the only way to know is to measure.
hThe harnessLive
Whatkya's hybrid agentic development system.
WhenApr 2026 →
HowAgents write the code, run the research, and ship real work; I run it the way I’d run any team.
WhyTo build for agents, you have to build with agents.
- Builds most of what kya ships, this site included.
- What it’s teaching me: the operating model changes before the org chart does.
IInkTreeLive beta · paused
WhatA reading app that replaces star ratings with a five-dimension taste graph.
WhenQ1 2026
HowFive dimensions (Craft · Mind · Heart · Pacing · Originality), feeling-based discovery, and a Goodreads importer that clusters editions into works. Shipped solo, with agents.
Why400 books rated on Goodreads, and four stars never once told me why I loved any of them.
- The first code I ever shipped to GitHub. Building it unlocked agentic development for me — InkTree is the reason kya exists. → inktree-beta.vercel.app
- Browse by feeling, not genre: “quiet books that hit hard,” “fast reads that changed my mind.”
- The build, in public: editions-vs-works, cover-art rabbit holes, 6,183 records clustered into 4,845 works. → the blog
- What it taught me: a love letter ships differently than a product — and shipping one is what made me a builder.
HHayStack ABIArchived
WhatAn alternative-data startup. The first build.
When2017 – 2019
HowCo-founded; led product and a team of six from nothing to working software and first revenue.
WhyTo find out if I could build something from zero.
- What it taught me: shipping v1 is a different sport than scaling one. I’ve played both.