What I look for when hiring
Context
I am currently hiring another data scientist to join our growing team at Wealthfront. During the process I have continued to calibrate what I look for when evaluating candidates. While participating on a panel discussion for Wharton FinTech in San Francisco earlier this week, I shared my short list with the students in the room. Someone joked that I should write a blog post to codify what I said. I agreed. Little did they know I was serious.
The following framework enumerates three critical dimensions that I try to assess when hiring, although it doesn’t describe everything I look for. For example, a candidate that doesn’t write code is not likely to make it very far as a data scientist, even if they cover the following bases. The three traits I list here capture what I think are some of the more difficult aspects to assess when hiring. They also generalize to functions other than data science, thereby making them potentially useful to more people. Although, I will add that they are more applicable to hiring for younger companies in the technology industry than for larger companies outside of it.
Mission
To outsiders, the trajectories of successful companies can seem preordained with the benefit of hindsight. For one glaring example, see Facebook’s user growth. In reality, this perception is usually far from the truth. Facebook’s post-IPO performance, when the company’s share price dropped by half, is a fantastic counterpoint. Startups and hyper-growth companies go through extreme ups and downs that the public never sees, at least until well after the fact (some view this as a good thing, others disagree). Many credit Facebook’s culture, namely its employees’ devotion to the company’s mission, with it’s ability to emerge vindicated from this period of doubt in the public markets.
Successes attract new members to the team while failures inevitably lead to churn. When a product launch flops or a TechCrunch reporter writes a mean article about your company, the people who are there for the wrong reasons are those most likely to abandon ship. Because I have experienced first hand how churn slows my team and our company down — in terms of both time spent rehiring and its effect on morale — identifying people who will stay on the rollercoaster until the end of the ride is top of mind when I hire for my team. The question I ask to try and get at this is simply, “why do you want to work at Wealthfront?”
Product
Young technology companies live and die by their product. This is because the product is what they build to do the job that their clients hire them for. Ideas for improving it can come from many places, including outside of the product team. Although I am not hiring for a product role, candidates with opinions and intuitions about Wealthfront’s product are high on my list of people to talk to. This is especially important for my team, given that the flavor of data science we practice involves spending plenty of time working with the product team to understand client behavior and iteratively improve our product.
Being customer-focused allows you to be more pioneering — Jeff Bezos
I have found this to be a useful way to screen for candidates that want to work at Wealthfront specifically — those hunting with a proverbial sniper rifle rather than a shotgun. That said, I make concessions here for candidates that are coming right out of school because as of today there is a financial hurdle to opening an account with Wealthfront. However, I do expect candidates to at least go through our sign up flow and watch some YouTube videos or ask friends to get a better sense for the product. The success of our company hinges on our product, and I want to work with other people who recognize this and make decisions everyday that will have a positive impact on the product and our clients. To get at this dimension, I ask candidates what they would improve about our product and how they would go about testing its effect if they were a data scientist working with our product team.
Learning
Teaching oneself new things is a necessity for people who want to succeed in startups and hyper growth companies. Because these organizations are doing a lot of things for the first time (perhaps for the first time by anyone ever) there is often no internal or external resource to use as an analog for what they are trying to achieve. This means teaching yourself about a given domain or problem and using what you have learned to reason from lower level principles about the right course of action. I have found there are three types of people when split along how they view this state of affairs: a handful of people see this as a negative, most see it as a necessary tradeoff for the other benefits of working at a young company, and a small number view it a compelling reason to work with you. I want to work with the latter.
Look for people who view the need to teach themselves new things as a feature and not a bug
I look for people who I believe can go to work on a problem and make tradeoffs and decisions with limited oversight. This is doubly important for data science teams because our field is so interdisciplinary. Whether you come from statistics, engineering, product or some other background, there will always be plenty of dimensions along which you can improve as a data scientist. To assess this I look for Coursera and Udacity certificates, or simply just ask candidates to tell me about something they recently taught themselves.
Disclaimer
I’m not sure that these criteria are optimal but I feel they are squarely in the ballpark in terms of characteristics of successful team members at young companies. If you’re looking for a more parsimonious description, Wealthfront Founder and CEO, Andy Rachleff, put it this way: “The worst thing you can do is hire for the job. You need people with headroom; people who could run your team some day.” I hope this simple framework is useful for other hiring managers out there. As always, please tell me if you disagree or feel that anything I said needs clarifying.