Build a project you can explain without an internship.

Choose a decision, then choose the tool.

Start with a small question someone could make a decision about. A spreadsheet, notebook or simple prototype is enough if you can explain the work. A complex stack without a clear question gives an interviewer little to explore.

Use a class, independent or volunteer project label that matches what actually happened. A personal project is useful evidence of practice; it is not an internship and should not be presented as employment.

• Analytics: examine a synthetic event-registration funnel and explain one follow-up question.
• Product: compare two public signup flows and propose a testable improvement; do not invent user research.
• Software: build a small task board with clear error states and a few meaningful tests.
• AI: create a source-linked FAQ prototype and evaluate it on questions you wrote; mark unanswerable cases.
• Finance or consulting: build a synthetic cost scenario, state the assumptions and explain a recommendation rather than claiming a client result.

A complete small example: event check-ins.

This is an original fictional teaching example using synthetic data. It is not a real university project or an observed student outcome.

Give the project a clear stopping point.

Write the question and the smallest acceptable deliverable before you begin. Build one end-to-end result first. Add features only when they improve the explanation or test an assumption.

• First work session: define the audience, question, data source and one deliverable.
• Next: inspect the data and record exclusions, duplicates and missing values.
• Then: produce one result you can reproduce from the original input.
• Finally: write what the result means, what it cannot establish and what you would test next.

Make the work inspectable.

A useful project page explains what you did, includes a small readable result and tells another person how to reproduce it. Remove secrets and personal data before sharing. Use only data and assets you have permission to publish.

In an application, describe the specific work you completed. Do not copy the example figures into your résumé. Replace them with your own measured results, or omit numbers until you have them.