CMS — Campus Management System

I shipped the marketing site at turtil.co — designed solo, then walked it through to production with my dev team's guidance. The site is live. I deployed it myself.

The product the site sells — the Campus Management System itself — is still being built. I'm one of the designers on it. This page is honest about the state: marketing surface live, platform in progress, deeper case study lands when the platform does.

On the marketing site: solo. Brief → research → strategy → iterate → prototype → test → frontend in Next.js + Tailwind → deploy. Devs helped when I got stuck on the build pipeline; I deployed it myself with their guidance. Turtil management gave me the room to learn deployment by doing it, and that's shaped how I work now.

On the in-progress platform: same loop, but as part of the team. I take briefs from marketing, run the research, ship prototypes, test the output, deliver the frontend. The dev team owns the backend, auth, integrations. Claude is my co-pilot through the whole loop — design, code, the words on the screen. Not a traditional designer anymore. Haven't forgotten where I came from.

I don't own the whole platform. Three surfaces are mine. Each one frames a different design question. Where I have an answer, it's below; where I don't, I've named the open thread instead of pretending.

1. 01

   Key modules across the platform

   Day-to-day screens for the modules I'm responsible for — the surfaces school staff and students will live in. Right now the work is mapping what gets designed first against what schools actually do on day one. Specifics stay under wraps until launch, but the principle is: design the registrar's Monday before the dean's quarterly review.

2. 02

   Onboarding AI

   First-day-on-the-job teammate, not a chatbot. When a new institution logs in, an AI reads their data and walks staff through setup — the imports, the role mapping, the first three weeks of operating rhythm. The open question I'm working: how much should the AI act on its own, and how much should it ask before acting? Trust beats automation. Erring toward asking, even when it's slower.

3. 03

   Campus assistant AI

   An assistant inside the platform — staff and students ask anything (schedules, fees, library, attendance) and get a grounded answer plus a one-tap action. We picked grounded answers with citations over free-form chat early, because wrong answers are worse than no answer in a school context. A teacher who can't trust the assistant on attendance numbers won't use it for anything.