Decide what each page has to prove
Every project answers one question for a hiring manager: what was broken, what I built, and what changed.
AI-ASSISTED BUILD / Personal project
This site: designed, written, and coded with AI, then refined through my own review and feedback.
SEE THE WORK
AI drafts fast. The edits are where the judgment shows.
Goal: a recruiter outside education should see what I built before they see where I built it.
Student information within reach Replacing paper with a system of record
Access to funded student services An intake-to-payment workflow across vendors
A path from grant funding to new schools Turning a $24.67M federal award into an executable program
Clear responsibilities across the school Decision rights and policy architecture for a new organization
School reviews that lead to action Portfolio reviews that turn findings into follow-through
The actual title revisions made to this site, shown in tracked-changes format. The before and after titles are real; the margin notes summarize the reasoning behind each change; the document view is a recreation.
THE STARTING POINT
I had strong work spread across résumés, case studies, and memory, and no single place that showed how I think. I also had no design or engineering team, which is the situation most operators are in when they need a tool built.
THE WORK
Every project answers one question for a hiring manager: what was broken, what I built, and what changed.
Recreate the real tools, from trackers to handbooks to apps, with fictional data so nothing private is exposed.
Check numbers against source documents, credit shared work, and cut anything I couldn’t defend in an interview.
Revise titles, results, and visuals in passes, the same way I would with a team.
HOW AI FIT IN
ChatGPT and Claude wrote the code and helped draft and restructure the content. I set the goal, chose what each page needed to prove, cut anything I couldn’t stand behind, and pushed revisions until the site reflected the work accurately. The demos use only fictional records. The site is its own proof: this is how I work with AI on anything.
THE RESULT
Project recreations use fictional data. Where numbers appear, they come from source records and are credited to the teams involved.