Most Python CVs are weak in the same fixable ways, which is good news: a couple of hours of editing can move you from "ignored" to "let's talk". Here is how.
A hiring manager gives you a few seconds
On the first pass nobody reads your CV — they skim it. In those few seconds they are checking three things: do you use the stack we use, have you shipped something real, and are you roughly the right level. Everything on page one should answer one of those.
The structure that works
- A two-line summary at the top. Who you are, your strongest stack, and the kind of role you want. No "passionate, results-driven team player."
- A skills line, honestly tiered. Separate what you use daily from what you have touched. Lying here gets exposed in the first interview.
- Experience as outcomes, not duties. Not "responsible for the backend" but "rebuilt the matching service, cut p95 latency from 800ms to 120ms."
- One or two real projects with a link to code or a live thing. For juniors this often matters more than the work history.
Lines that quietly kill your chances
- A wall of twenty technologies with no sense of depth. It reads as noise.
- "Familiar with" used five times. Pick a clearer word or cut it.
- No numbers anywhere. Engineering is measurable; a CV with zero metrics looks like one that has nothing to measure.
- A two-page essay. One page is plenty for most people under ten years in.
Tailor the stack words to the post
If a job asks for Django, Postgres and Celery, those exact words should appear on your CV — assuming they are true. Automated filters and skimming humans both pattern-match. This is not gaming the system; it is making the truth easy to find.
On PyHires your match with each offer takes into account your skills and the skills the offer asks for, so being precise and honest about what you actually know helps you more than padding does — padding just surfaces you for roles you would not enjoy.
The final test
Read your CV as if you were the person hiring. Can you tell, in under ten seconds, what this person is good at and whether they fit? If not, cut until you can. Then point it at the open roles and apply to the three that fit best, not the thirty that almost do.