Coinbase is a global remotely-based cryptocurrency exchange that enables users to buy, sell, and trade cryptocurrencies like Bitcoin and Ethereum. Currently, the Coinbase app is the company’s core product for retail traders, but they also offer crypto-related products, like USD Coin and solutions for institutional traders.

With the company’s 73 million users, data scientists at Coinbase hold an important role in making sense of all this streaming data through building user Lifetime Value models and analyzing user retention.

Coinbase data scientists have a number of responsibilities, which vary by team and job function. Data scientists at Coinbase may be responsible for:

• Measuring business performance and developing core metrics
• Building dashboards to track and understand KPIs
• Performing deep analyses and building models to understand customer behavior
• Extracting key insights that drive product decisions
• Performing data-driven analysis to drive actionable recommendations
• Synthesizing analyses to create compelling stories and visualizations
• Prototyping new analytics and machine learning models
• Engaging in cross-functional collaboration to drive new automations and reporting

Coinbase is big on culture. Therefore, desired qualities in Coinbase data science candidates are resilience and an affinity for learning, with interest in new technologies such as DeFi, NFTs, and Web 3.0. These data scientists must also be comfortable with communication and thrive in a startup-like environment.

Although required skills vary by role, in general, the company looks for:

• 5+ years relevant experience
• Evidence of core cultural values: clear communication, positive energy, continuous learning, and efficient execution
• Understanding of and experience in applying statistical concepts
• Experience in data analyses using SQL
• Experience in at least one programming language (e.g. R, Python, Java, Ruby, Scala/Spark, or Perl).
• Being able to independently create plans for analytics projects and build collaboration within teams
• BA / BS degree or equivalent practical experience

Coinbase has different stages of the interview process, where they ask candidates various data science interview questions. It generally includes a recruiter screen, a technical screen (including a SQL coding test), a take-home challenge, and ‘on-site’ rounds, which are remote since Coinbase has no physical headquarters.

Coinbase data science interviews are conducted remotely, and the process is similar to how they conduct engineering interviews. This process includes:

Initial HR Call- The first step is a short call with a Coinbase recruiter about the role. This call is used to determine if you’re the right fit for the role. Expect behavioral questions like:

• Why Coinbase?
• Explain your communication style.
• Do you have any questions about the role?

Technical Coding Screen- Coinbase technical screens are rigorous, and for data science roles, they typically focus on advanced SQL questions. Coinbase looks for “evidence you are able to produce production-grade code” and you will be assessed on the “end-result, as well as how you got the result.” The best way to prepare would be by studying:

• Advanced SQL questions
• Window functions
• Subqueries
• Query optimization

Coinbase Take-home Challenge - Coinbase data scientists must complete a 4-6 hour take-home assignment, and you typically have one week to finish the task (may be offered compensation). Take-homes tend to consist of root-cause analysis projects, and commonly incorporate skills like Python, SQL, and pandas.

Remote Interviews - The “on-site” is a series of one-on-one interviews, covering behavioral and technical questions. In general, you can expect questions about:

• Python data manipulation
• Algorithms (build an algorithm from scratch)
• SQL
• Behavioral questions

One helpful tip: Coinbase “bar raisers” sit in on interviews. Bar raisers are essentially a third party within the company that may or may not work on the data science team. Bar raisers have veto power on any hiring decision, so it’s critical to impress during any bar raiser round.

Specifically, Coinbase bar raisers look for candidates who:

• Inspire and impart knowledge
• Possess superior technical skills
• Could add value right away
• Are enjoyable to be around

There aren’t specific tips for passing bar raiser rounds at Coinbase. The best course of action would be to practice as many mock interviews as possible to simulate the interview experience and build your confidence.

Here are some sample Coinbase data science practice problems:

Write a query to get the number of customers that were upsold by purchasing additional products.

This medium-level SQL question is similar to what you might see in the SQL technical screen. You can solve this question with three sub-queries, but optimizing the query is key. You can do that with the HAVING clause.

Given a dataset of 600,000 credit card transactions, build a fraud detection model.

With this question, the most important step, after looking at all of these credit card transactions, is determining how we can feature engineer which data points are fraudulent transactions as our response variable. Once we have determined a high confidence for fraud, then we can build a model and extract features.

Name the two types of records that are present in the blockchain database.

Although crypto and blockchain knowledge isn’t essential for Coinbase jobs, having a basic understanding does help. With a question like this, mention both block records and transactional records. As a data scientist, you’ll likely be working a lot with transactional records.

Additional Coinbase Interview Questions