emusoiGitHub

Joneskim

/ˈdʒoʊnzkɪm/JONES-kimnounengineer, New York

  1. 1.Someone who loves physics and building things.
  2. 2.A software engineer at Lido, working on AI agents, data extraction and intelligent workflows.
  3. 3.A builder at emusoi, of tools for people who love building things, and build a lot of them: mia, buni and wazo.
  4. 4.informal A supporter of Arsenal, champions of England.

Origin Tanzania; UWC East Africa (IB, 2019–21); the University of Chicago (B.S. Physics, M.S. Computer Science, 2025). Shaped by ClassHook, the Vieregg Lab, GridLink and Haldune.

Scroll down for the entries

See also

buninounTypeScript, 2026

An open design file, and the tools for any coding agent to design in it.

Why Agents build interfaces straight in code, so the design never exists anywhere you can look at it, change it or hand back. It ends up only in the code.

buni gives the agent an open design file to work in, through the command line or MCP, with skills that teach it design. You watch the design take shape in the browser, and then the agent builds from the design it made.

ClassHookproper noun2022

Clips from popular TV shows and movies, for teachers to open a lesson with.

Joneskim was a software engineer intern here, in Ruby on Rails and React.

emusoiproper nounMaasai, discovery

A studio for building software tools, mostly for things that annoy Joneskim or that Joneskim finds useful, shared with the world.

Joneskim builds mia, buni and wazo here.

Why Most of what emusoi makes started as something that annoyed Joneskim, or something useful enough to keep. Each tool is built to be used every day first, then shared with the world.

Each takes a Swahili word for what it does; the studio’s own name is Maasai for discovery.

Halduneproper nounChicago

Software for post-incident debriefs and reports, so a team can learn from its past.

Joneskim was a software engineer here in 2025, building an unwritten knowledge platform.

Lidoproper nounNew York

An AI platform that extracts data from any document and puts it anywhere.

Joneskim has worked here since 2025.

mianounGo, 2026

Work on 100 things at once: a worktree, a session and a running app for every piece of work.

Why With coding agents, everyone works on a different part of the same app at once: a feature here, a fix there, an agent on each. One checkout can’t hold that. You stash, switch, reinstall and restart the dev server, the ports collide, and the laptop runs out of memory by the third app. Keeping track of what is where becomes the work.

mia gives every piece of work its own worktree, session and running app, and moves it to another machine when this one is full. Mia is Swahili for a hundred: work on a hundred things at once.

Vieregg Labproper nounUniversity of Chicago

A lab studying neutrino astrophysics and cosmology with radio and millimetre-wave detectors.

Joneskim was a research assistant here from 2023 to 2024, writing software for RNO-G, a neutrino observatory in Greenland.

wazonounTypeScript, 2025

Learn anything and keep it: notes that teach you, quiz you and find what you’re weak on. Your agents can write in them too.

Why Notes are where learning goes to sit. Material piles up, scatters across tools and never turns into anything you can practise against, especially when you’re teaching yourself with nobody checking the work.

wazo carries it round the whole loop: capture it, clarify it, practise, get tested, share it, come back to it later. Every note is a real file at a real path, so what you learn outlives the app, and the browser, the desktop app, the command line, Neovim and your agents all read and write the same files.