Following my previous blog post on how to land your first job as a Data Scientist, I received several questions from managers seeking advice on navigating the hiring process. In this article, I’m sharing some quick notes that might help you assess whether your hiring process is as rational as it should be.
Before we dive in, a small disclaimer: Since transitioning into the private data sector about seven years ago, I’ve been actively involved in hiring around 15 Data Scientist (DS) profiles. While this isn’t an extensive recruitment track record, these insights come from my personal experience.
Team Strengths/Weaknesses and Diversity
It’s common for DS teams to be understaffed when tackling projects. In such situations, it’s tempting to quickly hire the first person who seems “good enough,” is available, and motivated.
If you take this approach, you might overlook whether this person truly fits your needs. As Julie Zhuo, author of The Making of a Manager, says: It’s like being hungry at dinner time—anything will do—but if you had planned ahead, you’d likely eat something better.
Assess current and future needs
Take the time to identify upcoming projects, future opportunities, your team’s strengths/weaknesses in key skills, and how different personalities or experiences could enhance problem-solving diversity.
Through this exercise, you may realize that what you need isn’t a DS at all, it might be a Data Engineer (or two). Or perhaps your team already has enough white male economists. Maybe what you really need is a DS with more engineering skills (Type B) rather than analytical skills (Type A).
Make the Hiring Process Memorable
Design the kind of hiring process you’d want to go through yourself. If you’re working with HR personnel, forge an alliance where they bring expertise in recruitment and candidate engagement while you contribute technical knowledge about the role’s requirements.
Write the job description thoughtfully. Once you’ve identified what you’re looking for, craft a description that grabs candidates’ attention, not by overselling but by helping them see themselves in the role. A well-written description also equips recruiters with enough information to perform more effective filtering. I often repeat myself, but since DS roles are so poorly defined overall, make sure your job posting clearly outlines expected responsibilities.
Include a questionnaire alongside the CV submission. Ask about their knowledge of technologies, work methodologies they’ve used before, and their hands-on experience across various stages of a DS project pipeline: gathering data, cleaning it up, modeling it out, and even MLOps.
Additionally, find out how they’ve collaborated with other stakeholders: Have they translated business needs into analytical features? Integrated their work into other systems? Worked closely with engineering teams? Many DS projects fail due to poor coordination between teams and functions, experience here can be invaluable for your team.
Interviewing: During interviews:
- Repeatedly clarify what the position entails.
- Ask candidates how they perceive the role based on their background.
- Evaluate how well they understand what your company is looking for.
For example: Explain whether your team operates like artisans (where each member owns an entire project pipeline) or like an assembly line (where roles are specialized). If there’s an ongoing transition between these models, or if one approach defines your team, candidates should know upfront.
Provide context on how data is used within your company
During interviews:
- Explain how your organization approaches data.
- Share where your team fits within broader operations.
- Highlight its potential trajectory within decision-making processes or product development cycles.
For example: Is your team centralized or decentralized? How mature is decision-making based on data? Does the Chief Data Officer influence board-level decisions?
Encourage candidates to ask lots of questions
This signals curiosity, a critical trait for DS roles where helping others ask better questions often leads to better problem-solving outcomes.
Technical Test: Design tailored tests based on specific functions required for each role:
- If modeling isn’t central to someone’s duties, don’t include it in their test.
- Focus instead on tasks like handling outliers/missing values or drawing evidence-based conclusions versus relying solely on intuition.
When assigning technical tests:
- Avoid strict deadlines; flexibility matters, especially since many candidates complete these tasks outside regular working hours.
Selection Process: Structure candidate evaluations systematically by defining key indicators upfront and scoring consistently across all applicants. While consensus often emerges naturally during rankings discussions among interviewers—consider this dilemma carefully: Should decisions favor consensus or prioritize someone whom one team member strongly advocates?
Provide Feedback: Notify all candidates about their application status—even those who didn’t make it past initial stages deserve closure. For those advancing further (e.g., technical interviews), share constructive feedback about their performance so far, it builds goodwill and keeps doors open.
Remember, the community of data professionals tends to be tightly connected; avoid burning bridges unnecessarily! Candidates who ranked highly but weren’t hired could still fit future openings if needs align later down the line.
Lastly, involve other team members throughout profiling/hiring stages. not just managers! This broadens perspectives while giving everyone valuable exposure to processes they’ll likely oversee someday themselves.
Optimize Your Search
Segmenting searches via targeted mailing lists has worked well for me in past hires, for instance, if you are looking for talent in Spain: Spain AI, Comunidad de R Hispano or Pyladies are great examples worth exploring!
Leverage networks & trusted recommendations
Tap into personal networks heavily; ask trusted colleagues/ex-colleagues whose judgment aligns closely with yours when seeking referrals, especially useful when targeting junior profiles (e.g., reach out directly via academia/training institutions).
One final note: Attracting top talent requires offering strong workplace conditions, including engaging projects + reasonable access-to-data + career growth pathways! Demonstrate authenticity by showcasing real work through events/blog posts, it signals credibility clearly!