How to Screen a Backend Engineer Resume: Must-Haves, Red Flags and a Scoring Rubric
A backend engineer's CV is mostly nouns: languages, frameworks, databases. Screening it well means reading past the nouns for the verbs, what they built, ran and fixed, and for the scale it happened at. This guide is the checklist and the rubric.
- Read for evidence before matching keywords: a backend engineer CV should show what the candidate built, ran, closed or decided, with a result. The must-haves below are the four things the job cannot be done without.
- Must-haves: production experience with at least one of the languages the role uses, shown by systems shipped, not by a skills list; evidence of owning a service in production: incidents, on-call, monitoring, performance work. Nice-to-haves raise a candidate; they do not rescue one missing a must-have.
- Red flags are patterns, not single lines: a skills section with thirty technologies and a work history that used four, and every bullet starts with 'responsible for' and none with a result.
- Use the five-criterion rubric on the page, weighted so that production ownership counts more than job description match. It is written to be pasted into WadeCV as a role and scored automatically.
- Spend three minutes per CV with the rubric and write one line of evidence per criterion. Screening decisions with evidence attached can be explained to the candidate, to the hiring manager, and to anyone who audits them later.
Must-haves
- Production experience with at least one of the languages the role uses, shown by systems shipped, not by a skills list
- Evidence of owning a service in production: incidents, on-call, monitoring, performance work
- Relational database work beyond CRUD: schema design, indexing, query performance, migrations
- API design for other teams or external consumers, with versioning or compatibility concerns mentioned
Nice to have
- Familiarity with your cloud provider and deployment model
- Message queues, background jobs or event-driven work
- Testing practice described with specifics (what is tested, how, why)
- Contributions to shared libraries, internal tools or open source
Red flags
- A skills section with thirty technologies and a work history that used four
- Every bullet starts with 'responsible for' and none with a result
- No mention of anything breaking, ever
- Job titles inflate faster than the responsibilities under them
The rubric
Score each criterion from what the CV shows; weight 3 counts three times weight 1. In WadeCV, create the role with these criteria and every CV is scored this way automatically.
| Criterion | How to measure it | Weight |
|---|---|---|
| Job description match | How closely the candidate's stack, domain and scale match the role description as written | 2 |
| Years of relevant experience | Years building and running backend services, not total years in software | 2 |
| Production ownership | Evidence of running a service: incidents handled, monitoring, performance work, on-call | 3 |
| Data modelling and databases | Schema design, indexing, migrations, query performance described concretely | 2 |
| System design at scale | Largest system described in requests, data volume or users, and their role in it | 2 |
How to read the CV
Read the most recent role first and read it completely. The bullets that matter are the ones with a mechanism: 'reduced p95 latency from 800 ms to 120 ms by adding a covering index and a read replica' tells you the candidate understood the problem. 'Improved performance' tells you nothing.
Check the scale claims against the company. A candidate who 'handled millions of requests per day' at a ten-person startup may be telling the truth, but ask. Look for consistency between the size of the company, the size of the team and the size of the claims.
Do not screen out for a missing framework. Backend engineers move between frameworks in weeks; the things that take years are judgement about data, failure and change. Weight the rubric accordingly: production ownership above job description match.
Frequently asked questions
How much should a missing language on the CV count against a candidate?
Little, if the candidate has depth in a comparable language. A strong Go engineer will write good Python in a month. Weight the rubric towards ownership and data work, which do not transfer as fast.
What if the CV is all big-company experience and we are small?
Look for evidence they worked across the stack and owned outcomes, not just a component. Big-company engineers who mention on-call, deploys and cross-team work adapt well; those whose bullets are all about one service within a platform may not.
Should we use a coding test at the screening stage?
Not before a conversation. A short, real exercise after a screening call respects the candidate's time and gives you far more than a CV alone. At the CV stage, the rubric above is enough to decide whom to call.
Related guides
Screening backend engineer CVs this week?
Paste the job description into WadeCV, check the rubric it drafts, and upload the CVs. Each one is read against the rubric and scored with an explanation. Ten screenings are free; there is no subscription.
Screen your first CVs free