# Should every tech recruiter understand data science? **Authors:** Davide Facchinelli **Categories:** Data & AI **Tags:** ai-and-work **Last Updated:** 2026-04-13T14:10:32.700Z **Reading Time:** 6 min read --- ## Summary By Davide Facchinelli, maths lecturer at Albert School and former tech recruiter at Bending Spoons. Data science skills helped him recruit — but were they necessary? A practitioner's honest answer to a question most teams get wrong. --- *By Davide Facchinelli, Lecturer in Mathematics, Albert School Milan* It's 2026 — roughly 25 years into what may turn out to be one of the most consequential transformations in human history: the data and AI revolution. In a period like this, it is natural for almost everyone to ask: "Can I benefit from inserting AI and data science into my job?" I spent time as a tech recruiter at Bending Spoons after completing two master's degrees in pure mathematics and data science. I am now a lecturer in mathematics at Albert School. While working as a recruiter, I often asked myself how much I should lean on my technical background — and so did my colleagues and managers. This article is a reflection on those experiences and the opinion I formed. It is not a hard-science piece. We are talking about human behaviour, something no formal model has yet captured reliably — so it could not be. Before answering the question, we need to clarify what it even means. How could a tech recruiter use data science? There are two possible angles: understanding the roles and skills of the people you hire, and analysing the candidate data generated during the recruitment process itself — CVs, test results, interview signals — to improve hiring decisions over time. Both angles are real. The question is not whether data science brings advantages to a recruiter. It does. The question is whether those advantages justify making data science a standard competency for the role. --- ## Evaluating candidates: the human problem Let me start with what I believe is the most commonly misunderstood aspect of recruiting. Candidates — students especially — tend to think that the most important thing when applying for a job is demonstrating technical knowledge. That is not the case. The most important quality is the ability to learn and to collaborate. The ideal candidate is someone so genuinely interested in the work that it almost does not matter whether they can do it yet — they will figure it out. The harder, and more important, job of a recruiter is to assess exactly that: whether candidates are intrinsically motivated, whether they fit the team's values, whether they will thrive in that specific environment. The ability to conduct that kind of interview — and to read what it reveals — is the core competency of the role. This view is sometimes dismissed. Some argue that people should be hired purely on hard skills. I find that position unconvincing. The vast majority of people have an emotional relationship with their work. Pretending otherwise is not rigorous; it is naive. Recognising the human dimension and learning to evaluate it well is not a soft option — it is the work. --- ## Evaluating data: a serious cost-benefit question Now to the second angle: using data science to build better, more predictive hiring processes. There is a real scientific field for this — psychometrics, the discipline concerned with numerically measuring human traits such as intelligence, motivation, and personality. A well-designed recruiting framework could, in principle, correlate signals observed during the hiring process with post-hire performance, and improve over time. This is genuinely useful. But it does not require every recruiter to become a data scientist. A single person on the team with the right background, or a data scientist from another function, can build and maintain such a framework. The rest of the team benefits from the output without needing to generate it. And there is a deeper problem: building a robust, company-specific predictive model requires both data science expertise and a solid grounding in psychometrics. Psychometrics, as a science, is young. We have been working seriously on how to numerically measure human traits for roughly a century. Compare that to physics — one of the oldest sciences, with roughly 3,000 years of development behind it — and we still have very limited models of how the universe behaves. Humans are more complex than the universe. Asking a recruiting team to build a predictive model of human performance from scratch, without extensive historical data, is — to use a blunt analogy — like trying to land on the moon using physics as it existed before Newton. Not strictly impossible, but the probability of failure is very high. Does that mean we should not try? No. A process that improves good hire rates from 60% to 65% is a positive result worth pursuing. But the honest answer to "should we invest in building this ourselves or buy a validated framework?" is: almost always buy. Off-the-shelf psychometric tools are imperfect, but they are built on larger datasets and more rigorous methodology than most internal teams can replicate. The time and cost of building something better is, in most cases, not justified. --- ## The conclusion I reached My data science skills were useful in my recruiting role. But they were not necessary. My colleagues who lacked them asked an engineer for a couple of hours of support each week on technical elements and achieved the same results. I would not have advised any of them to spend time learning data science instead of deepening their interviewing practice or their knowledge of psychometrics. So: no, I do not think every tech recruiter should learn data science. Not because it is useless — it is not — but because I struggle to find a situation in which it is the best use of a recruiter's time, relative to what else they could be learning. It is a plus. It is rarely necessary. And in most cases, it is not even advisable as a priority. --- ## Two afterthoughts One practical note for anyone reading this while considering whether to learn data science: efficiency is not everything. If you are a recruiter who finds data science genuinely interesting, learn it. I was probably not at the peak of my time-efficiency ratio, but I enjoyed the work — and motivation produces better outputs than optimisation alone. One broader note: the fact that psychometrics is young and imprecise is not a reason to abandon the effort. Every science begins as a messy, loosely connected set of observations and gradually converges on models that fit reality better. Psychometrics will follow that path. The question is not whether to develop it, but whether each individual company's recruiting team is the right place to do so. --- *Davide Facchinelli holds master's degrees in pure mathematics from Université Pierre et Marie Curie and in data science from Sapienza Università di Roma. He worked as a technical recruiter and hiring manager at Bending Spoons before returning to academia. He is currently a Lecturer in Mathematics at Albert School's Milan campus.* ## Key Takeaways 1. Technical skills are not the primary hiring signal. The most predictive quality in a candidate is the drive to learn and the capacity to collaborate — not existing technical knowledge. Recruiters who optimise for the latter are measuring the wrong thing. 2. Data-driven recruiting is valuable; it doesn't require every recruiter to be a data scientist. A single person with the right background, or an off-the-shelf psychometric framework, can generate the same analytical value without pulling the whole team into a domain that isn't their core competency. 3. Psychometrics is a young science — treat it accordingly. We have been trying to numerically measure human traits for roughly a century. The models are improving, but expecting company-built prediction tools to outperform validated commercial frameworks is a high-risk bet in most contexts. 4. The right question is not "is it useful?" but "is it the best use of this person's time?" Almost any skill adds marginal value. The more important question is whether that investment yields more than the alternatives — and for most recruiters, deepening interviewing practice or domain knowledge wins. 5. Passion changes the calculus. Efficiency frameworks assume indifference between learning options. If a recruiter genuinely enjoys data science, that motivation compounds into better work — and that changes the answer entirely. --- *Article from [Albert's Deep Dive](https://deepdive.albertschool.com) - Albert School's Journal*