Topic 6 — Screening and interviewing

Goal: Assess candidates fairly and effectively in recruiter screens.

Lesson 6.1 — What the screen is actually for

Yara has Devon Asante on the calendar for 11:00. He's a senior backend engineer she spent two weeks coaxing out of a "happy where I am" reply, and the night before, she's quietly terrified. She's never written a line of production code. What if he asks her something technical and she freezes? What if he can tell, in thirty seconds, that the recruiter doesn't know a database from a doorknob?

Here's what calms her down once Imani, her mentor, spells it out: that's not what the call is for.

The recruiter screen is a short conversation — usually 20 to 30 minutes — whose only job is to decide whether a candidate moves forward to the hiring team. You are not testing whether Devon can design a fault-tolerant queue. The technical interview does that, run by engineers who can tell a database from a doorknob. Your call sits one step earlier and answers a smaller question: is this person worth an hour of Marcus's engineers' time? The screen doesn't measure how good an engineer is — it measures whether it's worth finding out.

That reframe is the whole topic. Once Yara stops trying to be a junior engineer for thirty minutes and starts being a sharp, warm filter, the dread lifts. She did exactly this for three years as an SDR, qualifying leads before handing them to closers. A recruiter screen is the same muscle, pointed at people instead of deals.

Lesson 6.2 — The four things you're checking

So if she's not grading his code, what is she listening for? Four things, and she can name them before the call even starts.

Rough fit. Does Devon's experience and level roughly match what Marcus opened the req for (Topic 3)? Not "is he brilliant" but "is he in the right ballpark." A senior backend role needs someone who's owned backend systems at scale, not a front-end engineer two years out of a bootcamp. Yara doesn't grade the depth; she catches the obvious mismatch before it costs five rounds.

Interest and motivation. Is he actually curious about this role, or just taking calls? "What are you looking for in your next move?" tells her more than any résumé line. If Devon wants to go deep on distributed systems and the role is mostly maintenance, that's a fit problem no salary fixes.

Logistics. The unglamorous stuff that quietly kills offers later: availability and start date, location and remote setup, work authorization (can he legally work here, does he need sponsorship?), notice period, and — the big one — salary expectations. Surfacing a comp mismatch now beats discovering it after five interviews and a verbal offer.

Communication. Can he explain his work clearly to a non-engineer? He'll be talking to Marcus, to product, to teammates. A basic read on how he comes across is squarely job-relevant.

And the part Yara's sales background already gets: the screen is a two-way sell. While she assesses Devon, she's also exciting him about Northwind — the robots, the Series B trajectory, the chance to scale a team from 40 to 75. A great candidate can walk away from a screen that felt like an interrogation, so she makes it feel like a door opening.

Lesson 6.3 — Asking questions that actually reveal something

Yara's first-ever screen, week two, went like this: "Do you have experience with backend systems?" "Yes." "Great. Are you a team player?" "Definitely." Twelve minutes, eight yes-answers, and she hung up knowing nothing. Imani listened to the recording and gave her one rule that changed everything.

Ask open, behavioral questions — ones that make the candidate show you instead of agree with you. Not "Are you a team player?" but "Tell me about a project you're proud of, and what you specifically did on it." Not "Do you like backend work?" but "What are you looking for in your next role?" Open questions hand the candidate a blank canvas; yes/no questions hand them a box to check.

Then the second rule, which Yara loves because it's pure sales-discovery: listen more than you talk. Her instinct, especially nervous, is to fill silence by pitching. But the person talking is the person revealing themselves, so a good screen runs maybe 70% candidate, 30% recruiter.

And the third, the one that separates a screener from a stenographer: probe vague claims for specifics. When Devon says "I led the migration to the new service," Yara doesn't nod and move on. She asks, gently, "When you say you led it — what did you actually do?" Sometimes "led" means he architected the whole thing and rallied four engineers. Sometimes it means he attended the meetings. The follow-up is how she finds out, without ever needing to understand the migration itself.

You don't have to understand the work to ask a good question about it. "What did you actually do?" works on any answer.

Her tech literacy from Topic 3 helps — enough to ask sensible follow-ups and notice when something sounds off — but the skill doing the heavy lifting is plain human curiosity, applied with discipline.

Lesson 6.4 — Structure and notes, so you can compare fairly

After a busy week Yara had screened six candidates for Marcus's backend req. He asked, "How does the third one compare to the first?" She opened her mouth and realized she couldn't remember. They'd blurred — different questions, different order, different mood — and now she had nothing to compare.

The fix is a structured screen: the same core questions for every candidate applying to the same role. Not a rigid script that kills the warmth, but a backbone. Yara keeps three or four anchor questions per req (relevant experience, motivation, a specific past project, logistics), asks them of everyone, and lets the conversation breathe around them. Now when Marcus asks who's stronger, she's comparing apples to apples.

The second half is notes. Yara writes clear notes during or right after every call, straight into the ATS (the applicant tracking system, the recruiting database, covered in Topic 9), usually on a simple scorecard: a few job-relevant fields and a recommendation. Memory is a liar across twenty candidates in a week. Written notes let her hand the hiring team an accurate, comparable summary instead of a vibe.

Structure plus notes do something beyond convenience: they make the process fairer, which is where the next lesson lives.

Lesson 6.5 — Fairness, bias, and the law you can't ignore

Two résumés cross Yara's desk for the same role. Same skills, same years. One name she can pronounce easily; one she can't. She notices a flicker of preference for the first — and stops, because she's been taught to. That flicker is unconscious bias, and a recruiter who gatekeeps for a whole engineering org cannot let it drive.

The defense is concrete. Judge on job-relevant criteria — skills, experience, ability to do the role — and be suspicious of "culture fit" when it's really a gut feeling. "I just didn't click with them" is often bias wearing a respectable coat. "They couldn't give a specific example of owning a backend system" is a defensible reason. Be consistent (that's why Lesson 6.4 matters) and check your assumptions: would I read this résumé the same way if the name or background were different? Don't reject on superficial signals — the strongest candidate rarely has the most polished résumé.

This is not only ethics. It is law. In the US, discrimination on protected characteristics is prohibited under Title VII of the Civil Rights Act (race, color, sex, religion, national origin), the ADA (disability), and the ADEA (age 40+), enforced by the EEOC. A biased screen is a legal liability for Northwind, with Yara's name on the notes.

Two more legal realities every modern recruiter must know:

Salary questions. Over 20 US states — including California, New York, Colorado, Washington, Massachusetts, and Illinois — now ban asking about a candidate's salary history. You generally can ask about salary expectations ("What are you looking for, comp-wise?"), and many states and cities now also require posting a pay range. So Yara asks "What are your expectations?" or shares Northwind's band — never "What do you currently make?" (The rules vary by jurisdiction and keep moving — a handful of localities have stricter wrinkles, and bills have been floated to limit even expected-range questions — so when in doubt she checks with Pri in People Ops before she leans on a script.)

AI screening tools. Everywhere, and risky. A 2024 University of Washington study found leading AI résumé screeners favored white-associated names about 85% of the time and disfavored Black male candidates in nearly every comparison. Laws are catching up: NYC Local Law 144 requires an annual independent bias audit of automated hiring tools and candidate notice. The employer stays liable for the outcome, so the tool never gets to be the excuse — human review on consistent, job-relevant criteria stays Yara's job.

Worked example — Yara screens Devon

The 11:00 call, done right.

Yara opens warm: two minutes on Northwind, the robots, why the role exists, selling from the first breath. Then she hands Devon the canvas: "Tell me about a recent project you're proud of, and what you did on it." He talks for six minutes about a service migration. When he says "I led it," she probes: "What did that look like day to day — architecting, coordinating, both?" Turns out he designed it and mentored two juniors through it. Strong, specific, real. She writes the actual detail into the scorecard rather than a bare "good."

She works her four anchor questions — the same ones she asks every backend candidate — so Marcus can compare cleanly. On motivation, Devon wants depth in distributed systems; the role has plenty, so that's a green light.

Logistics: authorized to work, no sponsorship; eight weeks' notice; open to the hybrid setup. Comp: she asks "What are your expectations?" — never what he earns now — and he names a number inside the posted band. No five-interview surprise waiting.

She catches herself, too. Devon's polished and easy to like, so she makes sure her "yes" rests on the specific migration example and the clean fit, not on charm.

After the call she writes a four-line summary: strong relevant experience, genuine interest, expectations within band, communicates clearly — recommend advancing to the technical round. Marcus reads it in thirty seconds and books the next step. That summary, not her memory, is the product of a good screen.