Enterprise seller based out of Chicago. Local-first AI enthusiast. Former college ballplayer, restaurant owner, avid golf trip attender, and the drawings are mine. Socials are down at the bottom if you want to connect.
Private things stay on my own hardware. The cloud still does the hard thinking — it just never sees the parts that are mine.
About
I'm an enterprise account executive at Lenovo, based in Chicago, covering Chicago and the New York metro. Nine years of selling technology to large organizations, mostly to people who have been sold to badly before and can tell immediately when it is happening again.
Somewhere in there I started building the tools I kept wishing existed. It got out of hand. There are now a couple hundred small programs running on a machine at home — they watch my calendar, index what I read, sort my inbox, run models I fine-tuned myself, and tell me when one of them has broken.
The rule I build by is that private things stay on my own hardware. A small local model reads anything sensitive and hands the cloud a summary instead of the document. The cloud still does the hard thinking. It just never sees the parts that are mine.
The other rule is that a language model may propose but never commit. The thing that actually writes to disk, moves money, or sends a message is always plain deterministic code with a switch I can flip. Agents draft. Code decides.
The drawings on this page are mine too. Same pen, same habit — start somewhere and let the pattern find its own edges.
Elements
A curated slice of what runs at home. Not everything here is public, and some of it never will be.
A four-role sales department, staffed by agents
Turns the exhaust of everyday sales work into briefings. Four agent roles share one local data pond; a full account brief renders in under a second, offline, with no model call. Customer data never enters the repo — only aggregate counts do.
Replaced the smart-home app with something I control
A local hub and a small web app on my own network. The browser can only call an explicit list of allowed actions per device type, so it can never invoke an arbitrary one. Reads keep working even when writes are switched off.
A voice assistant that runs on my own hardware
Local speech-to-text and local text-to-speech, with a tiered risk gate deciding what it may do on its own and what needs me to say yes first. Money and work systems are excluded at the capability level, not by a prompt.
Person detection at the edge, on real hardware
An always-on board watches for people and publishes to the home hub over a local message bus. The video never leaves the device — only the fact that someone is there does.
A wiki that maintains itself, drafted by a local model
Every morning an open-weight model reads what landed in my notes the day before, finds the gaps, and drafts the wiki pages that should change. A frontier model only reviews the draft. Promoting a page is a separate deterministic step with its own checks, so a bad draft can never quietly become the record.
Routes each job by how sensitive its data is, not just what it costs
Every headless agent call falls through a chain of models. What decides the route is the sensitivity of the data in the prompt — private work stays local, everything else goes to the best available model.
Recordings become notes without leaving the machine
An open-weight speech model transcribes the audio, a second small model cleans it up and titles it, and a third rewrites it into a proper note overnight. Every step runs on my own hardware. There is no cloud fallback on purpose. If the local model is down, the note waits.
Which open models can actually take over a job
Runs every local model through the same five tests: routing, extraction, instruction following, long context, and code. No model judges another. The graders are plain functions, so the same answer always gets the same score. This is what decides what moves off the frontier model.
Local model reads the private part; cloud model does the heavy lifting
A small model running on my own machine reads anything sensitive and hands the cloud a sanitized spec instead of the document. The good reasoning still happens in the cloud — it just never sees the private bytes.
Local models earn a job overnight instead of being handed one
A candidate open-weight model runs beside the frontier model every night on the same task, and its output is graded against the real one. Score high for five straight nights and it is allowed to take over. Slip below the floor for two and it is reverted. A timer does the judging, not a person and not a model.
A voice assistant on a desk box, with no cloud in the loop
A small speaker box wakes on a phrase. My machine turns the speech into text, an open-weight model answers, and a local voice reads it back. It can search my notes because a read-only index sits next to it. Nothing I say leaves the house.
Models
Each one is earning a job a frontier model does today, on hardware I own.
So the cloud model never sees a private document
Blacks out personal information in my own paperwork before anything is sent anywhere. It retrains itself nightly and only promotes a new version when the new one leaks nothing and reads at least as well as the one it would replace. It is trained entirely on synthetic documents — my real ones are never used as training data.
Off-the-shelf, the base model caught about a third of what it needed to. Tuned, it caught all of it with nothing missed.
Three chores a frontier model does today, rehearsed nightly by a local one
Filing my inbox, sweeping my backlog, and writing the nightly session retro. An open-weight model does each job in shadow, in strict JSON that a validator checks row by row, and a grader scores it against the frontier model's real output. None of the three has earned the flip yet. That is the point of grading.
Consulting
I take on a small number of AI projects on the side, with a network of engineers I trust for the parts I don't build myself. Local-first setups, agents that run unattended, and the guardrails that keep them honest.
It is free right now. I would rather do a few proofs of concept and earn the experience with you before anyone is charged. When it starts costing money, you will hear it from me first.
Questions? Message me on LinkedIn or email kabyrd012@outlook.com.
The stack
Every automation I run, scored on how likely it is to break. The broken ones are counted too.
