The advice that a photo on your CV will get you filtered out by applicant tracking software is repeated constantly and it's wrong in an interesting way.
Parsers don't reject documents for containing images. They extract text and ignore everything else. Your photo is invisible to them.
What actually causes damage is the document you build once you've decided to include a photo. That's a different claim and it's worth separating, because the fix is different too.
What a parser actually does with your PDF
A PDF is not a document in the way you think about documents. It's a set of drawing instructions: put this glyph at these coordinates, in this font, at this size. There is no paragraph, no heading, no column. Those are visual impressions produced by where things land.
So text extraction is a reconstruction job. The parser reads the glyphs and their positions and tries to work out reading order.
For a single column of text, that's easy. Sort roughly by vertical position, then horizontal, and you get back what a human reads.
Then people put a photo in the top right and everything downstream of that decision gets harder.
The three layout patterns that actually cause damage
Two-column layouts. Overwhelmingly the biggest one. To make room for a photo, people move to a sidebar layout: contact details and skills in a narrow left column, experience on the right.
A naive extractor sorting by vertical position interleaves them. You get "Senior Engineer" then "Python" then "Acme Corp" then "German (fluent)" as one stream. Modern parsers do better, but "better" is not "reliably", and you've no idea which one is reading your file.
Single column parses correctly everywhere. It's less pretty and it survives.
Text inside the header or footer region. Some extractors treat those separately or drop them. Putting your name and contact details there, which is exactly where a photo-led design tends to push them, means the parser may end up with a CV that has no name on it.
Text rendered as image. The killer, and the one that genuinely does look like "the photo broke it". If you exported from a design tool and the name is part of a graphic, that text does not exist as far as extraction goes. Same if someone scans a printed CV: the whole thing becomes one image with no text layer at all.
Test for this in two seconds: open the PDF, select all, copy, paste into a text editor. Whatever appears is what the parser sees. If your name isn't in there, nothing else matters.
That test catches more real problems than any amount of advice about photos.
So should you include one
Depends entirely on where you're applying, and this is a geography question rather than a technology one.
In Germany, Austria and much of continental Europe, the photo is still the norm in classical applications, even though the law stopped requiring it years ago. Leaving it out reads as unusual.
In the US, UK, Canada and Australia, the opposite. Many employers actively prefer no photo, some ATS-driven processes strip them, and recruiters occasionally return CVs asking for a version without one, because a photo creates a discrimination-claim surface they'd rather not have.
The guidance for each market is genuinely different, and what to include on a CV photo varies more by country than by industry. If you're applying across borders, keep two versions of the file. It's the same document with one image removed.
If you do include one
Single column. Photo in the top right of the first page, aligned with your name, not overlapping any text. Nothing important in the header or footer region.
Portrait format, 4:5 or 3:4. Not square, not landscape. 600 to 900 pixels wide, which is enough for print without inflating the PDF.
And crucially, the photo goes in as an image, next to text that is still real text. The moment you flatten a region to make it look nice, you've deleted it from the parser's view.
The format conventions for CV photos are more specific than people expect, and getting them wrong is what makes a CV look homemade in a way that's hard to name.
If you want something that comes out in the right aspect ratio without a cropping session, AI CV photos produce portrait format directly, which saves the step where you crop a square LinkedIn photo into a shape it wasn't framed for.
The actual failure mode
Nobody's CV gets binned by a robot for having a face on it.
Plenty of CVs get parsed into nonsense because the person moved to a two-column layout to make the design work, and then a recruiter searching for "Kubernetes" didn't find them, because "Kubernetes" ended up interleaved into a sentence about their previous employer.
Select all, copy, paste. That's the whole test.
More from this series
- Most AI headshot comparisons are testing the wrong thing
- 5 LinkedIn mistakes that filter engineers out before anyone reads the CV
- Background removal without a green screen is an underdetermined problem. Here's what the maths is doing.
- Warum dein Handy-Bewerbungsfoto deine Nase größer macht, und wie du das löst
On BetterPic
Top comments (0)