Skill bars communicate nothing to a parser

A row of filled circles beside “Python” is a design element. It contains no characters. When the resume is parsed, the word “Python” survives and every claim about your proficiency in it does not.

What disappears

• Skill bars and rating dots. Four-fifths filled means nothing to a machine and, honestly, very little to a recruiter either — nobody knows what your scale is.
• Charts and infographics. A timeline drawn as a graphic, a pie chart of how you spend your week, a radar plot of competencies.
• Icons. Location pins, envelopes, phone handsets, company logos.
• Photos. No parsing value, and in the US, UK, and Australia they are conventionally discouraged.
• Text inside a shape or text box. Depending on the generator, this may or may not land in the text layer at all.

The deeper problem

These elements do more than fail to parse. They take up the space where the evidence should be.

A skill bar occupies a line that could have read Python — 6 years, primary language at Northgate; built the fulfilment forecasting service. That sentence parses, matches keyword searches, and tells a human something specific. The bar does none of the three.

The fix

State proficiency in words, and prefer evidence to self-assessment:

Skills: Python (6 yrs), SQL, dbt, Airflow, AWS (Lambda, S3, Redshift)

Or better, embedded in the role where you used them, so the claim carries context.

For anything currently drawn as a graphic, ask what it is asserting and write that assertion as a sentence. If it turns out the graphic was not asserting anything, you have found something to delete.

Decorative elements that are fine

Horizontal rules between sections, background shading behind a heading, and bold or coloured type all render as styling rather than as content. They cost nothing at parse time. Keep the visual polish there.