Table of Contents


Introduction

Square, Inc. is a financial service, merchant services aggregator, and mobile payment company. Founded in 2009 and based in San Francisco, California, the company develops and markets hardware and software payment products that combine merchant services and mobile payments into a single, easy-to-use platform.

Out of their many robust products and services, people are most likely to be familiar with the Square iPad point-of-sale system used in many businesses across the country.

Square generates billions of monthly transactional data which forms the basis of analysis at the company. Data science is at the very core of Square’s products and features, and its data scientists are entrenched within various teams in a dynamic capacity to interpret and discern the vast amount of data that Square generates daily.

The Data Science Role at Square

The data science role at Square varies greatly depending on the unique goals and priorities of each team and cuts across a wide range of data science and analytics concepts. Depending on the team and product features, the role may range from mildly-technical, business/financial analytics-focused to deploying more technically advanced machine learning/deep learning algorithms. Thus, the tools and skills required may also range from basic analytics to write code to deploy machine learning systems.

Required Skills

Square hires qualified individuals with 2+ years of relevant industry experience.

Other relevant qualifications include:

  • An advanced degree (M.S. or PhD) in Computer Science, Software Engineering, AI, ML, Applied Mathematics, Statistics, Economics, Physics, or a related technical or quantitative field.
  • 2+ years (5+ for senior or 10+ for a lead role) of relevant industry experience in data science or machine learning-focused roles.
  • Experience deploying machine learning (e.g. regression, ensemble methods, neural networks, etc.), and Statistical (Bayesian methods, experimental design, causal inference) solutions to solving complex business problems.
  • Proficiency in any of the following scripting languages (Python, Java, etc.).
  • Experience with Hive, Google Cloud Platform, Looker, Snowflake, GCP, and AWS.
  • Experience with building complex, scalable ETLs for a variety of different business and product use cases.
  • Technical expertise in building personalization, ranking or recommendation systems that scale, with a fundamental understanding of machine learning algorithms and statistics.

Data Science Teams at Square

The term data science at Square encompasses a wide scope of fields related to data science. Data science roles at Square will likely be categorized under the title of  data scientist, data analyst, machine learning specialist, or product analyst. The roles and functions may range from product-focused business analytics to more technical machine learning/deep learning tasks.

More specifically, the data science roles at Square may include one or more of the following team-specific responsibilities:

  • Trusted Identity Team: Develop algorithms and cross-functional analytics to help understand Square’s customers and determine if users are complying with Square’s terms and conditions under the law.
  • Capital: Develop analysis and build models that help drive originations and reduce losses for Square’s business loan products.
  • Bureau Organization: Develop real-time data infrastructure and build algorithms to personalize Square’s products and services marketing efforts while enabling robust decision making across the organization.
  • Growth Data Science Team: Leverage data and automation to help Square solve impactful business, marketing, and growth problems such as lifetime value forecasting, churn prediction, attribution modeling, causal inference, and more.
  • Customer Support Automation (Cash App): Build models that anticipate customer issues and deliver proactive in-app suggestions, use NLP to contextualize inquiries and respond instantly with relevant content, develop prioritization algorithms that improve efficiency, and apply the latest research to automate conversations with customers.
  • Risk (Cash App): Build machine learning models that detect fraudulent activity in real-time, develop new product features that drive down risk losses, experiment with state-of-the-art algorithms to decrease false positives, use any and every dataset at your disposal (including 3rd party data) to engineer new features for risk models, verify customer documents using OCR, and use biometric and device signals to detect malicious logins and account takeovers.
  • Compliance Team: Build and automate actionable reporting, define KPIs, build ETLs, and build dashboards for key compliance processes to improve the overall compliance infrastructure and platforms.
  • Embedded Product: Leverage engineering, analytics, and machine learning to empower data-driven decision making in the full life cycle of product development while working cross-functionally across many different team organizations

The Interview Process

Square follows a similar hiring process to other big tech companies, with the exception of not administering a take-home challenge. The interview process starts with an initial phone call from HR to discuss past relevant experience and expertise. After passing the initial screen, candidates proceed to a technical interview (one or two in some cases) with a hiring manager and data scientist. If successful, an onsite interview will then be scheduled. This interview consists of four to five one-on-one interview rounds with several team members and managers.

source

Initial Screen

The initial interview is a 30-minute non-technical phone screen with HR or a hiring manager. The interviewer will ask about your past relevant projects as a data scientist to determine if you are a good fit for the team.

Technical Screen

Square’s technical screen consists of one to two rounds (varies by team) of technical phone interviews with a hiring manager and another data scientist. This interview is collaborative and it is done via coderpad. The questions asked during this stage are usually easy to medium-level Interview Query SQL questions as well as some modeling done in either R or Python (depending on the candidate’s preference).

Example Questions:

  • Can you give me the top 5 transactions and the maximum date for each?
  • Explain what metrics we should use to evaluate a binary classification model?

The Onsite Interview

The onsite interview is the final stage in Square’s hiring process. This is a full-day interview consisting of six rounds that will cover experimental design, statistics and probability, machine learning theory, modeling concepts, and team/cultural fit. Candidates must be ready to write code on a whiteboard in both SQL and Python along with a pair programming exercise.

The Square Data Scientist onsite interview will most likely consist of:

  • A coding and algorithms interview involving pair programming
  • A data exploration interview (also involving lots of coding)
  • A machine learning interview that involves white-boarding and explaining some of the fundamentals of machine learning concepts
  • A statistics interview that involves testing your knowledge of basic statistical concepts
  • An analytics interview which involves metrics definitions and applications
  • A culture fit interview with a product manager to determine if you are a good fit for the team and the company as a whole.

Notes and Tips

  • Remember, the interview process aims to assess how you can apply analytics and machine learning concepts to solve business problems, develop new features, and improve the customer experience. Read up on basic statistics and probability concepts, experimental design and A/B testing and statistical models and practice questions on SQL and machine learning models.
  • Since you are likely going to be white-boarding, it may be useful to practice coding on a whiteboard before the interview. Also, Square uses a shared coding environment for most of its technical interviews as it allows the interviewer to assess your technical ability at a glance while simultaneously getting a feel for your specific process. If you are new to pair programming interviews, we recommend reading up on the concept. It can be incredibly helpful to practice communicating your thought process and coding decisions out loud while writing code.
  • It helps to understand what it means to work at a financial technology company. This means understanding terms like revenue, profit, loss, etc.. and read up on how Square specifically makes money through their different capital, loan, and credit card reader services.

Square Data Science Interview Questions

  • Build a revenue model for the Square capital business.
  • Given a list of letters/strings with weights and a list of words, return the words with the highest value. Distinguish between capital and lowercase letters.
  • How do you test whether a new credit risk scoring model works? What data would you look at to understand the success of this model?
  • We are seeing exponential growth in the amount of users that are signing up for the Cash app. Draw a graph of what you think the number of weekly active users looks like.
  • Given a dataset of credit card transactions, build a fraud detection model.
  • Given an existing set of purchases, how do you predict the purchase of the next few items.