The GIMS gnome
Evan Brown
Solutions Architect  ·  Full-Stack Systems Builder  ·  M.S. Microbiology
Systems Builder · Lab Informatics · Technical Implementation
Highlands Ranch, Colorado · gims-project.net

Software for work too complicated to stay in spreadsheets.

Over eight years, I’ve turned regulated laboratory workflows into LIMS applications, automated reporting and validation systems, compliance infrastructure, and AI-enabled tools. My work spans requirements, architecture, data systems, deployment, and implementation with the people actually using the system.

Two of those systems are live: one has its first paying client, the other is in an active pharma pilot. Several more are tools I built because I needed them, and kept because they worked. All of it is below, with the real screens.

See the work ↓Read the resumegithub.com/BMA-Corgea ↗
The work real screens, not renders
The GIMS demo start screen: the gnome, the five grammar words, and the Start the walkthrough button
gims-demo.vercel.appthe real GIMS UI, driven by an in-browser mock
Information management · open core

GIMS

First paying clientOpen core on GitHub

A Python/SQL LIMS and database application that automates laboratory data entry and transformation. It replaces spreadsheet-driven workflows while supporting more convenient operation and cloud integration. You describe your lab as a grammar, and the engine generates the data model, the workbenches, the run logs, and the artifacts.

I [TEST] this [SAMPLE] on [HPLC] [for_COA]  →  a run that produces a Certificate of Analysis

The version everyone can use is gims-oss: the complete grammar engine and every workbench, editor, dashboard, guided tour, and skin, as a fully runnable single-user GIMS. Underneath it sits a full architectural refactor of a 100K+ line application: duplicated backend logic consolidated, SQL for the backend and React for the frontend, controls strengthened for audit readiness.

Case study →Open the demo ↗gims-oss on GitHub ↗
The Compliance Relay demo after a capture: the audit trail with a new record at the top and the gnome's bubble beside it
gims-compliance-relay-demo.vercel.appthe trail, one capture in
Compliance · 21 CFR Part 11

GIMS Compliance Relay

Active pharma pilot

A specialized vertical of GIMS, advanced through technical pre-sales. It provides the surrounding controls needed to use specific laboratory instrumentation in heavily regulated pharmaceutical environments without relying on handwritten records for organization and traceability: one append-only trail, electronic signatures, sealed exports an auditor can re-verify offline.

Case study →Open the demo ↗
A repo-tour page: a GitHub-shaped view of the repository with a four-chapter tour panel on the right
repo-tour/.repo-tour/pr-cb207c56.htmlrepo-tour touring one of its own pull requests · 14 stops through 12 files
Codebase tooling · tree-sitter, then a model

repo-tour and Build Tutorials

Be walked through a repository instead of opening it cold. Point it at a codebase you have never seen; it reads the whole tree, works out which few files actually carry it, and builds a guided tour, the system first, then the code, with the reasoning written by reading the lines it points at. Four of its five stages are deterministic and free. Exactly one spends tokens, and it runs last.

Build Tutorials is the inverse: a paint-by-numbers tutorial generated from a real repository that walks you through building it, one decision at a time, inside VS Code. The two share one core.

The suite →repo-tour ↗Build Tutorials ↗
The autoSQL picking screen: a pick on the left, the generated SQL, and two answer panes that disagree on one row, marked in coral
autoSQL/.autodev/evidence/T-2/04-disagree.pngthe disagreement state: SQL says 1, Python says 1e+300
SQL · an in-database query plane for GIMS

autoSQL

Public on GitHub

GIMS reshapes data by pulling it out of the store and running the transformation in Python. autoSQL moves that boundary: you pick what data and how to view, window, and transform it, and it compiles the pick into SQL that runs inside the database. The demo runs the same pick through both engines and shows every row where they disagree, because a silent wrong number is the one failure this project exists to avoid.

The project →autoSQL on GitHub ↗
Presence: a dark starfield with drifting cards, a Breathe card in the hand, and an interruption titled Inner board logic
meditationgamehost.vercel.app
A game · React

Presence

A small meditation toy about distraction, reaction, fixation, and the option to stop feeding any of them. A permanent Breathe card is always in your hand. Everything else appears on its own, asks for engagement, inflates a meaningless total, and expires.

The game →Play it ↗Source ↗
Also on the way: media-list, an organizer for all the stuff I’ve been meaning to get to, built to be more fun than the spreadsheet it replaces. Its cover-art wall is still being built; it gets a spot here when it exists.
How I work the AI is in the behavior, not the wallpaper

Every change to this site is a ticket.

I run my projects through a software factory built on Claude Code. A ticket goes intake, spec, a gate I hold, build, review, verify, a second gate I hold, ship. The models do the keystrokes. The ledger keeps the receipts. The page you are reading was built that way, and on the right is its ticket, as it stood when the page was last exported.

Underneath the factory is GUTS, which I created to organize my multi-agent projects: RAG, an MCP server, and agent-to-agent communication, used as the shared foundation for roughly a dozen tools and applications, ranging from launching applications to restoring an old inbox.

The whole factory, and the full trace →
portfolio-hub · .autodev/events.jsonlas of 2026-09-05 07:23 UTC · last 12 of 34 lines
06:31 T-1 goes back auto-review → build: auto-review FINDING (reviews/T-1.md): how-i-…
06:31 T-1 build: worker started (fable-5-1)
06:48 T-1 passes build (fable-5-1)
06:48 T-1 moves on: build → auto-review
06:48 T-1 build: worker reports back (fable-5-1) — /home/corgea/Desktop/Coding Projects/po…
06:48 T-1 auto-review: worker started (fable-5-1)
07:22 T-1 passes auto-review (fable-5-1)
07:22 T-1 moves on: auto-review → gate
07:22 T-1 gate merge cleared by policy:unattended
07:23 T-1 gate merge cleared by policy:unattended
07:23 T-1 passes gate (fable-5-1)
07:23 T-1 moves on: gate → verify
The lab where the software problems came from
Photo: Edward Jenner, Pexels · decoration, and labeled as such

Eight years in regulated laboratories, before and while building the software.

At Equilibrium Scientific I architected and delivered compliance systems for six-to-eight-figure laboratory operations across Colorado, New Mexico, South Dakota, New York, Akwesasne, Michigan, and the Caribbean, as the primary technical authority for validation architecture, reporting systems, accreditation strategy, and compliance automation across concurrent client engagements.

I rescued a Michigan ISO/IEC 17025 accreditation cycle by compressing approximately three months of unfinished preparation into two weeks. I designed and built a TypeScript-based Certificate of Analysis generation system using schema-driven reporting templates to absorb jurisdiction-specific reporting requirements without increasing operator burden. Before that I operated daily within an ISO 14644-classified clinical cleanroom at Eurofins, executing sterility workflows in a regulated LIMS environment, which is where I first identified the recurring mismatches between laboratory software workflows and bench-level execution.

Photo: Anna Shvets, Pexels
Photo: Brett Sayles, Pexels
Validation packages · accepted by state auditors
  • Microbiology✓
  • Potency✓
  • Residual solvents✓
  • Heavy metals✓
  • Mycotoxins✓
  • Additional assays✓
  • ISO/IEC 17025 · 21 CFR Part 11 · ISO 14644—
Also
  • Teams of 5–10 led through accreditation cycles✓
  • SOP systems sold, and the teams trained to run them✓
  • Bio-Techne: a multi-hour weekly reconciliation, automated to ~20 minutes✓