Most advice about tech hiring is written from inside the funnel. It focuses on the part candidates can see — the interviews — and tells them how to perform well in that part.
The harder problem is what happens before the interviews. For a single posting at a mid-to-large tech company, the rough shape of the funnel looks like this:
~250 resumes applied
↓ resume screen (seconds per resume)
~25 recruiter screen calls
↓
~10 technical phone screens
↓
~5 onsite or final-round interviews
↓
1 offer
The biggest dropoff is the first step, and that is the step candidates have the least information about.
Who actually reads the resume
Inside a company, the people involved in hiring play different roles, and treating them all the same is a mistake.
- The hiring manager defines the requirements and makes the final call. They are usually not visible until late in the process.
- The recruiter coordinates the pipeline and is, in practical terms, on the candidate’s side — they need the role filled.
- The sourcer finds candidates and brings them in. Their interest is encouraging but not a guarantee.
- An inbound screener, at larger companies, is the person who first looks at incoming applications. They scan, not read.
- The interview panel is trained and calibrated. Their job is fair assessment, not gatekeeping.
A lot of advice fails because it pretends a single “they” reads the resume. The screener and the hiring manager are looking for different signals.
What the screener can actually see
A screener usually looks at a resume for somewhere between five and twenty seconds. In that time they are scanning for signals that say “worth a recruiter call”:
- Recent role at a relevant company or scale.
- Concrete impact, not a list of responsibilities.
- A clean structure that does not require effort to parse.
- Match between the work history and the job description.
- Anything that flags an obvious blocker — visa needs, location mismatch, knockout questions answered the wrong way.
This is not the moment to be comprehensive. It is the moment to be legible.
The biggest lever is not the resume
The single biggest difference in pipeline outcomes is whether the application comes through a referral. Even a generic referral usually earns a more careful read. A strong referral — from someone who has actually worked with the candidate — often skips the screener entirely.
The practical implication is to ask before applying, not after. Once the resume is in the pipeline through the cold path, the referral bonus is usually no longer available, and the person being asked has less reason to spend social capital.
What the resume can actually argue
Inside the seconds it gets, a resume has one job: to make the case that this person fits this role. That is a different goal from showcasing every previous job.
The pattern that works is to write each bullet point as a small argument:
Accomplished {impact}, as measured by {number},
by doing {specific contribution}.
Most resumes do not contain numbers at all. A resume with a single concrete number per bullet point — team size, throughput, latency, cost, coverage, ticket volume, users — is already ahead of most of the stack.
The other lever is tailoring. Maintain a single long-form resume that contains everything. For each application, derive a shorter version that mirrors the language of the job description, surfaces the relevant experience, and removes anything that is not load-bearing for this role.
What is worth optimizing and what is not
Worth the time:
- Getting referrals before applying.
- Tailoring per job description.
- Replacing responsibilities with measurable outcomes.
- Keeping the layout boring and easy to scan.
- Applying early in the posting, not late.
Not worth the time:
- “ATS-optimized” templates and keyword stuffing.
- A summary section that repeats the rest of the resume in prose.
- Self-rated skill bars.
- Photos, dates of birth, marital status, or anything that triggers bias filters in markets where these are not expected.
- A second or third page that adds bulk without adding signal.
The funnel rewards clarity, not cleverness
The candidate cannot control how many people apply, who the screener is, or what mood they were in. What can be controlled is how legible the application is in the few seconds it gets.
That tends to favor the same things that make any technical writing useful: specific outcomes, plain structure, and the absence of decoration that distracts from the argument.