Projects

Practical problems, practical solutions

DCDS is a new practice, so most of what’s here comes from my own work: tools I’ve built for a nonprofit, my family, my teaching, and an online community. Each is labeled honestly. Some run every day, some are still being tested, and a couple I stopped on purpose.

Nonprofit · In use

Monthly grant reporting: from 10+ hours to one afternoon

The problem
A nonprofit board member was spending more than ten hours every month assembling grant reports.
The approach
She wanted to learn, not hand the job off. I showed her how to set up an AI assistant for the repetitive parts, and how to direct it and check its work.
The solution
A reusable AI skill that handles the tedious assembly, with her reviewing and finishing each report.
The result
About four hours, done in one sitting. She still doesn’t love grant reporting, but it’s no longer a multi-day chore. With more hands-on help it could likely go further; she chose to understand it herself.

More of my work, by the kind of problem it solves

These are my own projects, not client engagements. Each one shows an approach I’d bring to your organization.

Running daily

Reports that build themselves

Every morning, a set of scripts collects a snapshot of activity data for an online gaming community I’m part of, then posts a summary to the group: who was most active yesterday and this week, who leveled up, who joined or left. The same data later helped the group compare potential merger partners, with a dashboard to track activity as the two groups combined. A companion bot lets leaders schedule events and sends automatic reminders to the right people.

For your organization: the same pattern fits a weekly sales, attendance, or inventory report that currently gets rebuilt by hand.

In use

Right-sized tools

My family needed one shared view of a student’s online coursework, because two people couldn’t be logged into the course site at once. I exported the assignments and built a small private web app with drag-and-drop scheduling, so either of us can plan the week from any device. It’s deliberately simple. It solves that one problem and nothing more.

For your organization: not every problem needs a big system. Sometimes a small tool that does one job well is the right answer.

In testing

Working within a budget

With two of my students, I’m testing a setup where a paid AI plans the work and free models running locally do most of it, to see how far subscription costs can be cut.

For your organization: paid subscriptions aren’t always the answer. I look for options that fit your budget, and I’m upfront about the tradeoffs.

Proof of concept

Turning documents into something usable

I gathered six to eight years of my own course materials and used AI to build a profile of how I teach and write, which I now use when developing new courses.

Separately, I’m developing a process that takes a free, open textbook, classifies its problems, organizes them in a sortable dashboard by chapter, and generates import files for the course system, so instructors can give randomized homework without a paid platform.

For your organization: years of documents, procedures, and know-how can be organized so new staff, or a successor, can actually use them.

In testing

Software for small colleges

My largest project is an assessment platform for small colleges. Faculty collect data in the learning management system, an assessment director uploads it, and the platform creates follow-up tasks for reflecting on the data and producing gap reports. It works today and is in testing, with a commercial release planned.

For your organization: if your organization has a process that no off-the-shelf product handles well, a focused custom tool may be within reach.

Projects I stopped, and why

I built a battle calculator for a strategy game that accounted for roughly two thousand different player bonuses, and recruited friends to help test it. The math worked, but the interface was cumbersome, and if close friends wouldn’t push through it, wider demand probably wasn’t there. I shelved it.

I also built a fantasy football draft tool that allowed trading draft picks and could simulate practice drafts. It worked in testing, but my league decided to stay with its existing platform.

Knowing when to stop is part of the job. If something isn’t worth building, I’ll tell you.

Have a problem like these?

I’m glad to talk it through, even if it’s still just an idea.