The shape of the work
Most teams run the day across eight tabs and a spreadsheet nobody owns. I start with whatever’s slowing you down most — messy data, manual steps, tools that don’t talk — and work toward one place to run everything.
Services
-
Data cleanup & migration
I clean up your data so the numbers agree.
What an engagement looks like
Usually a week or two — landing on one set of numbers the whole team trusts.
- One clean place your data flows into, automatically
- Checks that catch bad data before it reaches a report
-
Workflow automation
The repetitive copy-paste work comes off your team’s plate.
-
Stack modernization
Replace the tools that cost more time than they save.
-
Systems integration
Get your existing tools talking to each other.
-
Unified operator portal
A single screen for the work, once everything underneath connects.
What an engagement looks like
A clickable mockup in a couple of days; a working first version in a few weeks.
- One screen, running on your real data — not another place to copy things into
- The things you do most, a click or two away
-
Decision support & ML
Forecasts, predictions, and flagging the odd thing out — once the data can support it.
Process
- 01
Watch
I sit with the people doing the work.
- 02
Map
The slow parts, and a plan with no padding.
- 03
Build
Designed and built by one person — me.
- 04
Hand off
You own it. Retainer optional.
Who I help
Non-profits
-
Executive Director
Stretched between operations, fundraising, and programs.
One clear view of how the org is doing, built on tidied-up data.
-
Development / Grants
Grant reports built by hand every quarter?
Connect your donor, accounting, and program tools so the reports build themselves.
Small & medium businesses
-
Founder / Owner
Every fix means one more tool to manage.
One simple tool that replaces several others.
-
Operations Lead
The "truth" of any number depends on who you ask.
Connect the systems and put one agreed number in front of everyone.
Mid-market companies
-
COO / Director of Operations
The team runs the day across too many tools.
One place to run it, starting with a single role and growing from there.
-
Data / Analytics Lead
Asked for ML before the data can support it?
Clean up the data first, then add models where they actually pay off.
About
Multiglass is run by Thomas Kavanagh, a Portland-based engineer with a decade in data science and ML. Discovery, design, and engineering by the same hands — mine.
Start with a quick call.
Thirty minutes — tell me where the work piles up.