Skip to content
View ariaxhan's full-sized avatar

Highlights

  • Pro

Block or report ariaxhan

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
ariaxhan/README.MD

Aria Han

AI consultant. I build autonomous AI systems that run companies end to end.

Not demos and not chatbots: systems of agents that plan, execute, check their own work, and escalate only the decisions a human should make.

I work in English and Korean, remotely, with clients in the US and Korea. A lot of my work sits on that line: teams in Seoul building for the US, US companies working with Korea, founders who think in one language and sell in the other. Nothing gets lost in translation because there isn't any — you get the same person on both sides.

I'm also a fractional CTO at an AI-first education company, where an autonomous system I architected supports the company's operations end to end. I own the architecture, the guardrails, and the exceptions.

Selected work

  • KERNEL — agent orchestration system: persistent memory, multi-agent coordination, validation gates, and an experiment engine for proving which rules actually work. Real external users; listed on the HOL plugin registry.
  • Nexus Office — the operations layer behind my consulting work: live agent threads, repositories, issues, and permission gates, running unattended on local hardware.
  • agentmailkit — Python package for scheduled, LLM-written email digests that run on your machine and send from your own inbox. Deterministic renderer; the model writes only the words.
  • Three free apps on the App Store (developer: CHAEUN HAN) — ModelMind, bite-sized AI mastery; Paper Rooms, every paper has an address; Civic Forges, build the city, live the life.
  • Writing — the novel The Correction, a speculative history on Amazon KDP, and the poetry collection notes on surviving eternity.

Work with me

Los Angeles · remote, US and Korea

Pinned Loading

  1. kernel-claude kernel-claude Public

    Claude Code learns from itself. Persistent memory, multi-agent orchestration, and a scientific experiment engine that proves which rules actually work.

    Python 13 1

  2. latent-diagnostics latent-diagnostics Public

    Representation-level analysis and supervision framework for large language models

    Jupyter Notebook

  3. memory-pool memory-pool Public

    Memory isn't a timeline.

    Svelte

  4. persist-os/vector-native persist-os/vector-native Public

    LLMs speaking their native language: vector operations, not English.

    Python 5 1

  5. experiments experiments Public

    Experiment engine for LLMs based on natural history specimens.

    Jupyter Notebook

  6. substrate substrate Public

    Taper-style generative gallery

    HTML 3