FanDuel Machine Learning Engineer (MLE) Interview Guide

FanDuel aims to give sports fans new and innovative ways to engage with their favorite teams, leagues, and games in a digital-first environment. Key products include:

• FanDuel - A fantasy sports app
• The FanDuel Sportsbook - An app for convenient sports betting
• FanDuel TV - A sports network serving live sports coverage and opportunities for fans to interact

Data is critical to FanDuel, and the company has embraced machine learning to build products that delight fans (while betting responsibly). As a Machine Learning Engineer (MLE) at FanDuel, sample projects may include game recommendation systems to improve user experience.

Are you a multi-faceted problem-solver? Are you excited about building machine learning-based models that scale? Are you interested in sports? If so, machine learning at FanDuel could be a good fit for you.

Machine Learning Engineering at FanDuel

FanDuel’s machine learning team sits within the Data organization.

There are three main Data groups—data architecture, data insights, and data engineering—which includes machine learning engineering.

Machine learning (ML) projects typically begin with an open business question. Data analysts and product managers work to understand whether existing tools can answer the question. If not, the project is passed to data science, which performs exploratory data analysis (EDA) to understand the problem better. If potential solutions require machine learning, the project is passed on to the machine learning team.

What does a FanDuel MLE do?

Once the project passes to the machine learning team, MLEs sketch out the end-to-end technical architecture:

• Sourcing the data.
• Building the data pipelines.
• Engineering model features.
• Building and training the model.
• Testing and model validation.
• Model deployment, monitoring, and maintenance.

The ML team is also responsible for building a robust infrastructure to support the model’s needs. MLEs own the service that encapsulates the model, so a big part of the job is stress testing the service by integrating it with the product, developing a support/rollout plan, ensuring that correct performance metrics are chosen and observable in production, as well as coming up with a plan to facilitate post-production experimentation if FanDuel ever needs to switch out a certain model or dataset.

Because MLEs at FanDuel are concerned with building production-ready models that scale, candidates will need more than just knowledge of core ML concepts. You must be a great programmer and effective problem-solver.

It is up to you, as an MLE at FanDuel, to ask questions like:

• What cloud resources do we have access to to support this project?
• What data do we have on hand?
• How is this data maintained?
• How often is it updated?
• What data pipelines support this data, and are these sufficient to feed a working model?
• How will we handle sensitive data?

MLEs must understand service-oriented and pipeline-oriented architecture because they are responsible for building the service that runs the model. They must also understand what it means to experiment with a model versus working with it at scale. You will constantly solve problems in data engineering, feature engineering, DevOps, infrastructure management, and more.

For example, a posting for a Machine Learning Engineer lists the following tasks:

• Using data transformation technologies, designing and implementing data pipelines required in the data warehouse and data lake in batch or real-time.
• Identifying, designing, and implementing internal process improvements: automating manual processes, optimizing data delivery, and re-designing infrastructure for greater scalability.
• Designing and deploying data models and views with large datasets that meet functional / non-functional business requirements.
• Delivering data integration solutions to downstream marketing and campaign software.
• Delivering test plans, monitoring, debugging, and technical documents as a part of the development cycle.
• Creating data tools for analytics and working with stakeholders across all departments to assist with data-related technical issues and supporting their data infrastructure needs.

What are the typical job requirements for a FanDuel MLE?

Excluding internships, candidates will typically need:

• Education: Bachelor’s degree or equivalent in computer science or engineering (or a similar field) or at least 1-3 years minimum of work experience.
• Specialized experience: Real-world data engineering experience. FanDuel looks for MLEs who know how to handle messy and incomplete data. There’s no expectation that you’re an expert in every machine learning model. Still, you should be a strong programmer (especially in Python and SQL) with experience building and using data pipelines, feature engineering, and building robust, production-ready models.

Here are a few of the basic requirements listed for the Machine Learning Engineer job mentioned above:

• Experience as a data engineer in a machine learning environment (recommendation systems, pattern recognition, data mining, or artificial intelligence).
• Hands-on experience with ML frameworks and libraries (Scikit-learn, Pytorch, Tensorflow, LightGBM, Keras), data structures, data modeling, and software architecture under scalability, correctness, and maintainability constraints.
• Processing event-based data using streaming technologies like AWS Kinesis and Kafka for ML pipelines.
• Working with other members of the ML platform team to support the delivery of additional project components (API interfaces and microservices).
• Designing and building data pipelines for production-level ML infrastructure using tools such as TFX, Kubernetes, Kubeflow Pipelines, and TensorFlow.
• Experience demonstrating technical leadership working with teams, owning projects, defining and setting technical direction for projects.

Finding salary benchmarks is challenging as FanDuel’s ML team is small right now, but per Levels.FYI, software engineers at Flutter (FanDuel’s parent company) make:

• $154K per year at the L3 level, with a $129K base, $17K in stock, and $7K in bonuses.
• $195K per year at the L4 level, with $156K base plus $29K in stock and $10K in bonuses.

Recommendations before you apply for FanDuel MLE roles

• Revamp your resume. Communication skills are critical at FanDuel, and your resume is the perfect way to demonstrate this from the beginning. Make sure you can tell a coherent, compelling story around all the experiences listed on your resume and why you’re an ideal candidate for FanDuel.
• Practice with mock interviews. Exponent's coaching services are your best friend. Don’t limit your pool of mock partners to other MLEs and peers in tech—grab a non-tech friend and describe the most recent project you spearheaded. Communicating effectively with your peers in engineering and non-technical collaborators will be critical for your growth on the job.
• Lean on your community. Take what you learned from your team research and connect with a few machine learning engineers or FanDuel data engineers on Exponent or LinkedIn and ask them about their experiences. They’ve gone through what you’re going through now and are great sources of information and support.

Interview Process

FanDuel’s interview loop is short and sweet. You’ll undergo:

• A phone screen with a recruiter.
• An onsite (or equivalent) consisting of four rounds, including a technical assessment, a behavioral/technical retrospective with the hiring manager, an ML system design interview, and a culture fit round.

Overall, the process is quick. Expect a final decision within a few weeks.

Recruiter Phone Screen

FanDuel’s recruiter screen is standard. The call typically lasts for 45 minutes to an hour, with some time reserved at the end for your questions. This call ensures both the candidate and the hiring team are aligned on expectations and general fit to not waste either party’s time.

Your recruiter will explain the role, describe the interview process in more detail, and ask questions to gauge whether you’re a potential match.

Prepare to:

• Describe your technical experience, emphasizing your familiarity with real-world data and deploying production-ready ML models.
• Answer questions about what motivates you and why FanDuel is a great fit.
• Describe what you’re looking for in an MLE role long-term and what you could bring to this small (but growing!) team.

Onsite (or Equivalent) Interview

If you pass the recruiter screen, you undergo four more interviews assessing various skills.

Expect to face:

• A technical retrospective/behavioral round with the hiring manager to review your resume and discuss your experience.
• A coding round split into two segments: Data processing and SQL.
• A system design round where an experienced data architect or MLE will ask you to design a scalable ML-based system.
• A culture fit round assessing collaboration and leadership skills.

Technical Retrospective & Behavioral

Your first interview at FanDuel will be with your hiring manager. Recent FanDuel hires describe it as a conversational blend of a traditional behavioral interview and a technical retrospective (or “retro”). You’ll deep-dive your resume and answer questions about your experience. The hiring manager aims to assess your technical experience and your fit with the ML team and FanDuel’s company culture.

FanDuel looks for candidates who are:

• Enthusiastic, hard-working, and collaborative.
• Strong problem-solvers who are willing to fail quickly and learn from mistakes.
• Able to identify and translate business requirements into technical requirements.
• Able to think holistically and design robust systems that scale.

Expect to discuss ML or data engineering projects end-to-end, including scoping requirements, building data pipelines and cleaning data, model training and testing, and deployment and maintenance. Be prepared to explain your technical decisions in detail, emphasizing your contribution (as opposed to the team’s) and its effect on the business.

Here’s a list of technologies, languages, and tools commonly used by FanDuel MLEs. If you have experience with these, highlight those skills in this interview.

• Programming Languages: Python, Pyspark, SparkSQL, SQL, and more. Infrastructure: Terraform
• Cloud: AWS EC2, ECR, RDS, Redshift, DynamoDB, Sagemaker, AWS Kubernetes, Redis, Kinesis, Lambda, etc.
• Feature Engineering: Tecton
• Data Engineering: Data transformations, pipelines, airflow, Kafka, Databricks, DBT, etc.
• Software Engineering Principles and Devops: Experience in a CI/CDbased environment, build/deploy tools like Buildkite or Github actions, experience with Git-based development, etc.

Urgency is essential to FanDuel (the online sports betting market is competitive and constantly evolving). You can expect questions about handling pressure, strict deadlines, conflicts, and/or challenging team dynamics. To ace these questions, describe how you handled these situations in the past and the critical lessons learned.

Coding Round / Technical Assessment

The technical assessment aims to simulate a task you’ll perform day-to-day rather than asking you to solve common coding problems.

You’ll spend about 30 minutes on an ML-oriented task (typically using Python, but you can choose another language) and 30 minutes using SQL to answer basic business questions.

Because real-world data is messy and data processing is critical to quality ML outputs, you’ll want to spend time cleaning and exploring the data first. A common pitfall is latching onto a trendy model and building without exploring the data thoroughly. Your interviewers will be watching your process - how you read in data, identify gaps or errors, and build an efficient solution. Explain your process and strive for quality, though perfect code isn’t required.

Next, you’ll spend another 30 minutes performing short SQL tasks. Practice table joins, group by functions, aggregations, and window functions so you can quickly answer simple business questions.

System Design and Architecture

The system design round s is where you’ll demonstrate your ability to think through building an end-to-end ML system. Your interviewer will assess your knowledge and thought process as you go.

Interviewers want to see you:

• Determine business requirements and translate them to technical requirements.
• Articulate a high-level solution using machine learning.
• Consider various components and model tradeoffs, and make appropriate choices.
• Design a robust infrastructure to handle incoming data and make design tradeoffs.
• Ensure your solution is scalable, serviceable, and easy to monitor and maintain through deployment.

You’ll have an hour to define requirements, build your system, and evaluate your output. We’ve included some topics to practice and a short framework for working through ML system design questions below.

Culture Fit

In the final culture fit round, you’ll meet with the hiring manager again. This round is primarily behavioral and conversational. The hiring manager wants to confirm that you are enthusiastic about FanDuel and will mesh with the small machine-learning team.

Instead of one-sided behavioral questions, expect to have a discussion covering:

• Your career goals.
• Your vision for ML at FanDuel.
• How you will be an asset.

To prep for this round, spend some time considering what drives you. How do your motivations align with FanDuel’s company culture and mission?

Top MLE Interview Questions

Data Processing and SQL

• Write an SQL query to return top players from each of two teams based on their single highest score.
• Write an SQL query that returns records from the (given) users table while excluding any duplicate email entries.
• You are given a table with varying distances from various cities. How do you find the average distance between each of the pairs of cities?
• Given an e-commerce database (with the following schema), write a query to fetch the product name and stock of the most recently purchased product.
• Given a players table, write an SQL query that returns the names, scores, and ranking of the 4th, 6th, and 11th-ranked players.

System Design and Architecture

Behavioral / Culture Fit

Technical (Retrospective) Questions

Sample Interview Questions

Technical Retrospective

The technical retrospective assesses technical knowledge and decision-making, but there is a strong behavioral component as well. Expect questions on gathering information before making decisions, the group dynamics, and how you balance stakeholder interests.

To prepare, consider the following for all major projects on your resume:

Context

• What was the context of this project? What was the problem to be solved, and who was it affecting?
• What were the business goals and justification for the overall project?
• What were the different pressures constraining the project? Was it a steep technical challenge or an urgent issue requiring quick action?

Technical Decisions

• What were the technical requirements? How were they decided?
• What resource constraints existed?
• What ML model/component/architecture choices did you make and why?
• How did you scale the model and service architecture?
• What tradeoffs did you make, and what happened? Would you have done anything differently?

Team Leadership

• Who were the key stakeholders?
• What was prioritized, and how?
• Who did you collaborate with, in what way, and what information did you gather?
• If you were a people manager, how did you divide work among your team?
• Did you need to protect your team’s time amid urgent stakeholder requests? How did it go?
• How did you keep stakeholders informed of progress?
• Were there any issues, miscommunications, or unforeseen risks? How did you manage these?
• What did you learn?

Then, practice telling your story in a concise, impact-oriented way. We recommend using the STAR framework for both behavioral interviews and the technical retrospective.

Are you looking for more practice?

Exponent has an interview prep course dedicated to technical retrospectives. It’s a good idea to practice live with a partner, as you’ll surely get several follow-ups. Exponent’s behavioral peer mocks run twice daily—check them out here.

Data Processing and SQL

The coding round will assess your ability to process data using a language of your choice and SQL.

Common data questions include:

• Build a simple data pipeline to accomplish a task given a dataset.
• Given a dataset, identify errors and clean the data so it’s ready to be fed into an ML model.
• Given a file, what ML model would you choose to accomplish a task?

The best way to prepare for coding questions is to get lots of practice. Begin your prep by reviewing common data structures such as:

• Graphs and Trees
• Arrays
• Linked Lists
• Stacks and Queues
• Hash Tables

Then, start coding. It helps to review data engineering fundamentals if you’re rusty. Check out KDnuggets’ curated list of free data engineering courses for helpful resources.

For the SQL interview, expect questions like:

• I have data in one database, but I need to migrate it to another—how would you do this?
• Write a window function to rank data in a database.

To ace the SQL interview, review:

• Querying Missing Values
• GROUP BY and HAVING functions
• Table Joins
• Aggregations
• Window Functions

You can review many standard SQL functions and watch mock interviews in Exponent’s SQL interview course.

Here are a few tips for success in coding interviews.

• Code in your best language. Don’t choose Haskell if you’re unfamiliar with these but want to impress. Interviewers want to see you at your best.
• Don’t be afraid to Google a helpful library or function. You’ll have full use of the internet to complete your task. No one expects you to have the entire Python documentation memorized.
• Articulate your thought process. Always!

ML System Design and Architecture

The system design interview assesses your technical knowledge and experience with the entire ML lifecycle and your ability to translate business requirements into technical requirements and build a quality service to support and deploy the model at scale.

Common prompts include:

• Design a fake news detection system.
• Design a landmark recognition system.

As you design, you’ll answer additional questions and follow-ups like “What are the tradeoffs at the service architecture level regarding building a batch training vs. real-time training system?” Interviewers aren’t looking for a perfect solution but are interested in how you think. You’ll be assessed on your ability to:

• Translate a business problem into a set of technical requirements.
• Choose a high-level solution using machine learning.
• Anticipate potential data issues and outline the data pipeline required.
• Design the service architecture such that your chosen model performs optimally and produces the required outputs while ensuring the end-user experience is positive.
• Ensure your solution is scalable, serviceable, and easy to monitor and maintain through deployment.

Follow this 5-step framework to ace your system design round at FanDuel:

• First, define the problem. Ask clarifying questions until you thoroughly understand the ML task and have concrete requirements.
• Then dive deeply into the data. Discuss data sourcing, data cleaning, how to handle different data types (for example, real-time vs. batch collected), and articulate how you’d build a robust data pipeline.
• Next, choose a model and outline the infrastructure required. Discuss how you would train your model and choose a set of performance metrics to evaluate the model.
• Then discuss model deployment, monitoring, and maintenance activities.
• Finally, review your design. Review your requirements again, suggest any changes you’d make given more information, summarize tradeoffs, and answer any questions.

Your CS and system design fundamentals may be rusty if you haven't interviewed recently. Check out Exponent’s Fundamentals of System Design course for a thorough refresher and real-world mock interviews.

Behavioral

You’ll get a few behavioral questions during your recruiter screen, the technical retro/behavioral round, and the culture fit round. Your hiring manager leads two of these, so prepare some questions for them as you practice articulating your own story.

To prepare for behavioral interviews, first review FanDuel’s mission and company values:

• Everything begins with the customer. They have real money on the line, so we take their fun seriously. Our revenue comes from customer delight—never the other way around.
• We are one team. Learning from the past helps us focus on the future. We act on behalf of the entire company, not just our team.
• Assume positive intent. Everyone at FanDuel is intelligent and hard-working. Even during disagreements, we remain calm and respectful of our colleagues' opinions.
• Anything is possible. We dream big and pride ourselves on being trailblazers. We benchmark against the best in the world, not just the industry.
• Simpler is always better. We break down the seemingly impossible into bite-size challenges. We solve problems through innovation so our customers don't have to.
• Win with integrity. We are relentlessly hungry to win but don't sacrifice our integrity. It’s the foundation on which our business is built.
• We say thank you. We recognize that achieving the highest standards requires relentless intensity from our people, so we always show appreciation to our teammates and partners.
• Own the outcome. Influencing results and making the journey sustainable are in our control. Trust is earned by being vocally self-critical and open to continuous improvement.
• Stay humble, stay hungry. Learning never stops whether you are a coach or an athlete. Our people are our strength. Learning, development, and leadership are our priorities.
• Challenge, then commit. Nobody's view is perfect. If you disagree, voice your concerns calmly and respectfully. But we commit to delivering the best possible outcome once we decide.

As you practice common behavioral questions, ask yourself:

• What do I do to ensure my actions align with my team and company?
• Why do I want to work at FanDuel? What about FanDuel’s vision resonates best with me?
• How have I shown leadership in my past projects and personal relationships?
• How have I been an effective communicator and collaborator? What has my experience taught me, and what will I bring to FanDuel?

There are no right answers to behavioral questions, but we recommend practicing enough to be succinct and straightforward but not robotic. Interviews want authenticity and passion here, so don’t be shy!

Be sure to practice these common behavioral questions at FanDuel:

Don’t forget to check out Exponent’s behavioral question bank. There are hundreds of questions and answers ready for practice.

Tips and Strategies

• Emphasize real-world application of ML. Experimenting with ML on a small scale vs. deploying it to solve business problems at scale are entirely different activities. If you have real-world experience, lean heavily on that. If not, always ensure that you’re suggesting models and infrastructure choices that scale well.
• Read the job description carefully. Make sure you understand the scope of responsibilities and what FanDuel is looking for for each open role. Check the list of technologies used at FanDuel and tailor your responses to showcase your experience with these.
• Communication is key. Asking for clarification is always better than proceeding with incorrect assumptions. Interviewers are there to help; don’t hesitate to ask to repeat or reframe the question.
• Solve out loud. Create visuals—they’ll help you clarify your thoughts and help your interviewer keep track of your process. Remember to answer all questions (both technical and behavioral) with the role in mind. Consider how your ideas and experience relate to what FanDuel does.
• Be yourself. Recent FanDuel hires stress that FanDuel has a fun, welcoming culture and that you should feel safe to be yourself in the interview. No one expects perfection. Try to have fun with it.

FAQs

• How many rounds of interviews does FanDuel have? Typically, FanDuel interviews consist of one recruiter screen plus an onsite loop (or equivalent) of four rounds.
• Are the interviews at FanDuel difficult? FanDuel interviews aren’t challenging if you properly prepare! Practice behavioral questions, ML-based system design questions, SQL, and data processing using your favorite language.
• How long does it take to get an offer? The timing of the FanDuel interview to offer varies widely.
• Does FanDuel negotiate on salary? You should always try to negotiate your salary depending on your experience.