Wiki

CV Screening

Archive-backed guide to how data CVs and resumes are screened: responsibilities, keywords, project evidence, recruiter calls, bias reduction, and ATS myths.

Related Wiki Pages

Job Search Hiring Job Descriptions Data Scientist Interview Roadmap Data Science Careers Data Analyst Careers

CV screening is the first hiring filter after sourcing or application. In the DataTalks.Club archive, recruiters and hiring managers use it to decide whether a candidate should enter the interview process. It sits between job descriptions and the later interview rounds described in Data Scientist Interview Roadmap.

Recruiters usually compare the CV or LinkedIn profile with the role the team actually needs. They check role fit, level, keywords, and project evidence. They also check communication and practical constraints. Candidate-side preparation belongs to Job Search. Employer-side funnel design belongs to Hiring, but the two topics meet here because the CV is where both sides first test the match.

Common Definition

The common definition across the archive is simple: CV screening is evidence matching. The screener asks whether the candidate’s written evidence matches the job, the team, and the expected level.

Alicja Notowska gives the recruiter-side version at 21:32 in Hiring Data Scientists and Analysts. She starts from the job description and the hiring-manager discussion. Then she checks profiles for matching keywords, experience, education, and concrete responsibilities. The strongest signal isn’t a title. It’s evidence of what the person personally did.

Oleg Novikov gives the candidate-side version at 18:28 and 25:51 in Data Science Interview Guide. He frames the CV as a landing page for the next hiring step. The reader should see the relevant contribution quickly, without unrelated detail hiding the match.

Luke Whipps adds the market-map version in Land Data Scientist Roles. At 7:35-11:59, he starts with role definition and candidate longlists. At 14:07-27:19, the CV screen checks industry and use-case fit. It also checks projects, business impact, and the candidate’s career story.

Guest Differences

Guests agree that the CV should prove fit, but they weight the signals differently by role and market.

Alicja’s data-science and analyst screening puts more weight on education when a team asks for research depth. At 27:10 in Hiring Data Scientists and Analysts, she separates research-heavy teams, where a PhD and papers may matter, from other teams. For those teams, a bachelor’s or master’s degree can be enough if the candidate shows the right work.

Nicolas Rassam weighs titles and degrees less for Data Engineering. At 31:16 and 50:45 in Hiring Data Engineers in Europe, he looks for SQL and Python. He also looks for real projects, outcomes, specific skills, and evidence that the candidate keeps learning. A candidate can come from BI or software engineering. They can also come from another data role if the project evidence is strong enough.

Luke gives design and positioning more weight than the other guests. At 14:07 in Land Data Scientist Roles, he says the first impression starts when the recruiter opens the CV. He still checks substance, but he treats formatting and information hierarchy as part of professional clarity.

Oleg is more explicit about ATS myths. At 55:17 in Data Science Interview Guide, he separates readable parsing from template-rejection myths. He treats automatic template rejection as a weaker explanation than candidates often assume. Candidates should still write for fast human scanning and simple software parsing.

Screening Signals

Personal contribution is the strongest signal. Alicja gives the recruiter version in Hiring Data Scientists and Analysts. At 22:13 and 28:41, she says a profile with only company names gives the recruiter little to evaluate. Titles and company descriptions aren’t enough either.

A useful entry says what the person did and what they owned. It also says whether they built models, pipelines, analyses, or products themselves.

Keywords help the profile appear in sourcing, but they need surrounding evidence. At 25:56, Alicja describes searches for machine learning and AI. She also uses ML, deep learning, algorithms, and role-specific must-haves. At 32:40, she warns that buzzwords without responsibilities can pass the first text match and then fail the first interview. That connects CV screening to Job Descriptions: the keyword should come from the job, but the CV must explain the work behind it.

For data engineering CVs, Nicolas looks for a smaller number of tools that the candidate truly used. At 31:16 in Hiring Data Engineers in Europe, he names SQL and Python as basics. Then he asks for the problem, the data, the tools, and the outcome. At 40:06, he says cloud and BI tools are transferable if the candidate understands how and why they’re used. Claiming expertise in many tools creates risk because interviewers may ask for details.

Level also changes the screen. At 26:38 in Hiring Data Engineers in Europe, Nicolas ties screening and assessment to junior, mid-level, and senior expectations. Senior candidates need clearer tradeoff reasoning around time, money, performance, and bottlenecks. They also need to explain system choices. Junior candidates need stronger evidence that they can learn, execute, and explain their work.

Layout matters because recruiters scan quickly. At 28:41 in Hiring Data Scientists and Analysts, Alicja says experienced candidates usually lead with work experience. Recent graduates may lead with education or projects. Clear dates with month and year reduce ambiguity. At 56:47 in Land Data Scientist Roles, Luke discusses country differences and a two-page guideline rather than a fixed universal length.

Personal details rarely help the screen. At 44:38 in Data Science Interview Guide, Oleg recommends removing age and photo when they’re not needed. He also recommends removing address and marital status. Alicja gives a similar warning around 28:41-35:49 in Hiring Data Scientists and Analysts: irrelevant personal information doesn’t prove job fit and can introduce bias.

Portfolio Evidence

Portfolio evidence matters most when the candidate lacks direct role history or when the project proves a skill the CV claims. For data science candidates, Oleg opens Data Science Interview Guide with a project-as-differentiator example at 2:42. At 45:46, he recommends projects, synthetic data, and blogging for PhD-to-industry candidates who need visible applied evidence.

Luke evaluates portfolios through role fit. At 19:50 and 25:04 in Land Data Scientist Roles, he looks for links between tech stack and project. He also looks for links between use case and business impact. That’s the portfolio version of a good CV bullet. It tells the reader why the work mattered, not only which library appeared in the notebook.

Nicolas gives the data engineering standard at 54:25 and 55:53 in Hiring Data Engineers in Europe. Strong examples include first data pipelines and business-specific datasets. Privacy work and data deletion systems also stand out. So do projects the candidate can explain to a nontechnical recruiter before going deeper with engineers.

A GitHub link can help, but he doesn’t treat public GitHub as mandatory. The candidate still has to explain their exact part of the work.

Those portfolio links belong in the CV-screening graph: Data Engineering Portfolio Projects covers pipeline evidence, and Machine Learning Portfolio Projects covers applied ML evidence. Open Source Portfolio Evidence covers public work such as issues, pull requests, documentation updates, and maintained tools.

Recruiter Screen After the CV

The first recruiter call usually tests the claims that survived the CV screen. Alicja says at 36:08 in Hiring Data Scientists and Analysts that recruiter screens rarely go deep technically. They clarify responsibilities, gaps, and motivation. They also cover salary expectations, notice period, active hiring conversations, and communication. For senior data scientists, she may ask the candidate to explain complex work in nontechnical language.

Nicolas makes the same point for data engineering at 44:35 in Hiring Data Engineers in Europe. Candidates should know what the company does, why they’re talking, and how to describe their projects to a nontechnical person. At 48:13, he contrasts this with broad, untargeted applications where candidates can’t explain why the company or product interests them.

The CV therefore has two jobs. It gets the candidate into the call, and it gives the recruiter a script for the first questions. Weak bullets create vague questions. Specific bullets create useful conversations about career transition, role fit, technical depth, and next interview steps.