AdaptiLab Blog

Emulating an Industry Standard Hiring Process

Posted by James Mahoney on Feb 18, 2020 1:17:00 PM

In the article "How to Consistently Hire Remarkable Data Scientists", Jeremy Stanley, former VP of Data Science at Instacart and Chief Data Scientist at Sailthru, describes the process his team at Sailthru used to effectively hire data scientists. Stanley writes that "most managers worry that they regularly lose amazing talent because their process is so long and cumbersome." Not only do these long processes cause potential hires to drop, but it also wastes time for your internal data scientists to do all the interviewing. He then writes that the "ongoing effort that hiring requires can easily consume 20% or more of a data science team's time." That's a lot of time wasted, compounded upon by the fact that data science talent is already so scarce at companies.

Stanley proposes three steps to follow to build an efficient data science hiring process. We will first summarize these steps, and then go into greater detail on how AdaptiLab helps you mimic this process.

  1. Candidate Generation: We want this process to have a predictable funnel of talent. The key is figuring out how to attract new candidates and then making their application process as fluid as possible.
  2. Base Technical Competency Evaluation: A hard technical barrier to separate candidates who are technically proficient from those who are not. This should be easy to administer, but difficult to pass.
  3. Company Specific Technical Competency Evaluation: Create a challenge that is curated specifically for your company, and preferably uses data that emulates the type of data the company works with. You can either give this as a take-home or an on-site data project.
  4. Onsite Interview: Determining both general cultural fit for the company and the team-specific fit of the individual.

Now that you have a general idea of what an effective data science hiring process should involve, we will go into it in more depth. The next parts of this post will outline how to take these elements of the hiring process and implement them in your current one (with the help of AdaptiLab's products).

Step 1: Establishing A Solid Candidate Generation Funnel

Getting candidates can be a whole process in and of itself. With so many companies investing in machine learning and data science talent, it can be difficult to compete, especially against large corporations with ubiquitous brand recognition. Because of this, there are things you must do to set yourself apart from other companies that your potential hires may be considering. Firstly, you must establish an online presence. Whether it is with blogs, social media posts, or a combination of both, giving candidates an idea of what they can expect at your company, and why your company would be an interesting place to work at, is key. One element we have found that is vital to attracting candidates, and determining their fit, is company culture. Try to make your company culture evident in your online presence.

Second, make it easy to apply to your job posting online. By using some form of applicant tracking system, it's easy to track online applications. When combined with our online Code Screen product, applicants can apply to jobs and be assessed with no effort on your end. By using this approach, you may get applicants who you would otherwise overlook, but actually have the technical abilities to succeed as a data scientist or ML engineer.

Finally, finding pre-vetted candidates is another way to fill the pipeline. One source for these types of candidates is AdaptiLab's online marketplace.

In our online marketplace, we have candidates who have passed our comprehensive evaluation process: Code Screen, background phone call, and Take Home Project. Candidates have been pre-vetted by AdaptiLab and it's completely free to sign up and get access to the portal.

Step 2: Initial Technical Competency Screening

The initial technical screen should be accessible to any candidates that apply, but set to a high technical bar. This ensures that candidates who you dedicate time and resources to interviewing have the necessary technical skills to be a competent data scientist. While some people use technical phone screens to solve this, a more efficient and less tedious way of screening talent is by using our Code Screen product. After a candidate applies to your job position, you can send the candidate a screening challenge through our platform. The candidate will receive an email with the link to the screening challenge hosted on the AdaptiLab platform.

A screening challenge generally takes between 1-2 hours, depending on what exact questions/areas you want to focus on (we have hundreds of questions to test candidates on specific competencies necessary for the position). The Code Screen tests candidates on the five core competencies of data science: data processing, data analysis, feature engineering, model development, and general algorithms.

After taking the screening challenge, candidates receive immediate feedback on their score and what areas they excelled at and/or need to improve on. We've found that feedback for candidates is something they value highly, and with our screening tool it is automatically delivered to candidates post test. On the hiring manager's end, you will see a more detailed score report with comparisons of correct answers and the candidate's answer. Given that AdaptiLab's product has been built by industry experts, the questions and their answers are in-depth and complete. Thousands of candidates have gone through the process and a large majority prefer it to traditional phone screens.

Step 3: Company Specific Technical Competency Phase

With the previous step, we narrowed our candidate funnel and determined who has the base necessary competencies in data science and machine learning. Our Take Home Project gets even more detailed than the Code Screen, with more in-depth questions. The questions in our Take Home Projects are built around data that emulates the type of data and problems encountered at your company. By doing this, we evaluate candidates on the exact tasks you will need them to complete as employees. You can choose to have them either solely submit their findings (which we grade for you with our network of industry data science graders) or both submit the findings and present them at an on-site. Bringing candidates on-site to present their findings allows your hiring managers to assess the candidate's communication ability and business acumen, a crucial skill for data scientists.

Our custom Take Home Projects are graded by industry data scientists. The specialists share feedback and competency review in a comprehensive report that we provide to you along with the candidate's raw submission.

Step 4: Onsite Interview

No matter which path you decide on for the Take Home Project, it is imperative that you bring your candidate to an on-site. There is no substitute for in-person interaction to assess how the candidate will work with your team and your company's culture. By making the first elements of the hiring process as standardized and streamlined as possible, you can be objective and thoughtful when assessing a candidate's overall cultural fit.

AdaptiLab wants to help you build a great data science team while reclaiming your interviewing time and delighting your candidates with an enjoyable and relevant interview process. We've worked with seed stage startups up to Fortune 1000s, and we're confident our wide range of products can help you optimize your interviewing funnel. Reach out to, or sign up for a demo with us. We'd love to learn about your current data science hiring struggles, and help you figure out an effective solution.

Subscribe to AdaptiLab's Blog and join thousands of hiring managers and HR professionals!

You May Also Like