← Back to Home

Day 2: Building Craft & Defining Growth

Two separate threads today – one about building a tool, one about building a career.


Craft: An Internal Design Growth Program

Started with a context dump on Midwestern’s internal design process: the 5-phase design thinking framework, AI-augmented tooling across Zoom, Claude, Figma, Notion, Cursor, and GitHub. The problem to solve: build an internal alternative to PIPs and growth plans – something positive, self-directed, and useful during non-billable time.

The output is Craft – a single HTML file. No login, no tracking, no dependencies. Seven pre-populated learning buckets:

  1. Research & Discovery
  2. UX & Systems Thinking
  3. Visual Craft
  4. Prototyping & Motion
  5. AI & Tooling
  6. Communication & Storytelling
  7. Front-End Literacy

Each bucket contains curated resources, time estimates, and editorial notes on why the material matters. A ghost card holds space for team-defined buckets down the road.

Visual Refinement

Went through two rounds of iteration after the initial build:

Round 1 – standardized typography to DM Sans, tightened border radii, opened up spacing.

Round 2 – full light-mode restyle to match midwestern.com. Cool lavender-gray background, white card surfaces, lime highlight treatment, hairline grid dividers, and a dark “Why it matters” panel pulled from their site’s callout section pattern.


Leapsome Goal Development

In a separate conversation, built out a full FY 26-27 professional development plan: phased quarterly timeline, SMART goals, free learning resources, and a weekly time budget. After seeing the first draft, pulled it back intentionally – less ambition on the code side, more tinkering than building.

From there, expanded into two additional Leapsome goal tracks:

The final pass tightened everything into Leapsome-ready language: summary statements, key results, and a reusable context block for future AI chats.

What I Learned

Writing goals for a platform like Leapsome forces a useful constraint – vague intentions don’t survive the format. The elicitation round before writing made a significant difference in output quality. Getting the right inputs before generating anything is a pattern worth repeating.