AI consultant
Building since 2024
Aria Han
AI for work, AI for humans.
Workflows, systems, and everything else to go from “we need AI” to “we have AI.”
Where I tend to be useful
I like technically ambitious work with thoughtful people.
The projects are chapters, not trophies
How do people learn AI? How do people read research? How does AI remember? How do teams work with agents every day? The projects are different answers to questions that keep coming back.
Thirteen projects, grouped by the questions underneath them: learning AI, reading research, keeping context, coordinating agents, preserving evidence, and making daily work less likely to evaporate.
- ModelMindAI education felt backwards, so I built the course I wanted people to have first.
- Paper RoomsSafari tabs ate my research life, so I built a library.
- our4cutsAn iPad and a browser are all you need for a photo booth.
- HeyContextContext kept disappearing, so we built a workspace around memory.
- HeyContentCreators had context everywhere and nowhere, so we tried to bring it into one place.
- Brink MindMental health tools ignored the body, so Brink Mind brought heart data into the room.
- KERNELClaude Code kept starting over, so I built memory, rules, and receipts around it.
- llm-benchLeaderboards were not answering my questions, so I made tests that did.
- the-agent-libraryPrompt collections kept rotting, so I turned repeated workflows into portable skills.
- model-familiarity-engineI wanted to know what a model had earned across a working relationship, not where it ranked.
- metabrainMemory gets noisy unless it has to prove itself.
- SubstrateI wanted to see what an unattended creative pipeline would become if it kept going.
- latent-diagnosticsCorrect answers were not enough. I wanted to know whether the model computed something real.
AI consultant · Blink Build Studios. My current work focuses on internal AI workflows. Outside that engagement, I also review and repair AI products for founders and independent builders. I follow new models, tools, methods, and research closely, then update the work when they make a better approach practical.
Still active in the open: KERNEL, my memory-and-rules layer for Claude Code; llm-bench, the 21-test model benchmark; and the daily Substrate pipeline that ships one agent-made artwork a day. Also active but mostly invisible: the daily automation system that sends me research digests, keeps the vaults alive, and occasionally turns the machinery into poetry.
The full timeline →Notes from inside the work
Essays from the questions I keep circling: agents, memory, tools, language, and what all of this is doing to us.
This is a record of recurring frictions. Learning AI felt backwards. Research papers kept disappearing into tabs. Context kept getting lost. Work kept restarting every morning.
I have built products with users and pitch decks. I have also built free apps with no ads, no paywalls, and no plan to extract anything from anyone.
The difference is not the business model.
The difference is whether the system helps people keep learning, keep context, keep evidence, or keep a conversation alive.
I'm motivated by meaning, and by the question underneath all of it: how to use AI to make humans more human.
So I build for continuity: memory that accumulates, tools that disappear into the background, names that make systems easier to think with, and AI that does the work people should not, leaving the human part more intact.
Building continuity in a world that keeps fragmenting.
Everything within reach
Each object opens a room. Pick one up.