Meta Data Engineer Interview Guide

Meta Data Engineer Introduction

Meta (aka Facebook’s) Data Engineer’s interview process includes SQL and Python coding, ETL jobs including both batch processing and streaming (with PySpark), data modeling, product sense, and behavioral/culture fit questions.

Facebook’s data engineer is a highly paid, well-respected role spread across their entire company and business units. Because of its massive amount of customer base (billions of monthly active users), data engineers at Facebook have plenty of opportunities to use some of the most advanced large-scale data engineering technologies, many of which are homebrew such as HIVE and Presto.

Meta Data Engineer Salary Guide

Salary data sourced from instamentor’s mentees and cross-checked over the internet.

Every situation is different, e.g., the offer also depends on your interview performance, job location and the urgency of filling that role. Therefore, this data is for your reference only and may not apply to everyone.

Meta Data Engineer required skills

Here is a baseline curriculum our mentors can help you prepare. We will adjust it based on your level of skills and the hiring team.

SQL (More META/FACEBOOK SQL Interview Questions)

• Depends on the group, but is usually focused on querying vs insert/update
• JOINS (including Self-Joins with complex ON conditions)
• Window Functions (with an understanding of how to implement these using self-joins)

Python

• Main python data structures (base types, lists, dictionaries)
• Function creation / handling
• Decorators

Batch processing

• Basically, all data is provided in chunks with tools such as Spark or Hadoop

Streaming

• Data provided as it comes, need to maintain state in various forms
• How to aggregate with the state in SQL (ie, new data vs historical table)

Data Modeling

• Ability to design schema for a given scenario (ie, rideshare service, food delivery service, etc)
• Build queries to access data in the model
• Update model/queries as more information are required
• FB tends to lean on star schema models

Product Sense

• Engagement
• Growth
• Retention
• Churn

Behavioral

• What’s the first thing you would do…?
• How have you handled conflict…?
• What’s a decision you made you were proud of…?
• *STAR* method

Meta Data Engineer Interview Case Study

Jake, one of our latest mentees, recently accepted a great data engineer job offer from Facebook.

Here is what the whole process looks like:

TLDR

Candidate: Jake

How it gets started: applied in May, a friend did an internal referral

Job level: E5

Year of Experience: 5–10

Degree: M.S & B.S. in CS

Offer: Yes

TC: ~450K USD

Location: Menlo Park, CA

Interview process: 2 months

Preparation: 2 months

Has a job: yes

Decide to join: Yes

Technical screen round 1:

Python and SQL

Technical screen round 2:

SQL

Final round (met 4 people)

Python

SQL

Data Modeling

Leadership/Behavioral questions

Sample Python questions (4 questions to be completed in one hour)

1. Fill the None values with the previous none None value

[1, None, 1, 2, None] → [1, 1, 1, 2, 2]

Tips: You have to pay attention if there are consecutive Nones.

2. Write a function to return a list with words that don’t have a match (case sensitive) between two strings

(“Facebook is an awesome place”, “Facebook Is an AWESOME place”) → [“is”, “Is”, “awesome”, “AWESOME”]

3. Write a function that counts the frequency of a character

(‘missisipi’, ‘s’) → 3

4. Write a function that returns the key of the nth largest value in a dictionary

Example: {‘a’: 1, ‘b’: 2, ‘c’: 100, ‘d’: 30}

n : 2 (2nd largest value)

output: ‘d’

Sample SQL questions

1. Percentage of paid customers who bought both product A and product B
2. Percentage of sales attributed by promotion on the first day and last day of promotions.
3. For each product A, find the top 5 other products that people also bought.