Software Engineer, Machine Learning - Fraud

As a Software Engineer, Machine Learning - Fraud at DoorDash, you will conceptualize, design, implement, and validate ML solutions specifically aimed at preventing and mitigating fraud. Additionally, you'll collaborate with cross-functional teams to shape the product roadmap, mentor junior engineers, and lead efforts that drive significant changes in our fraud detection technology.

To apply for the position of Software Engineer, Machine Learning - Fraud at DoorDash, you should ideally have a Master’s or PhD in a quantitative field such as Statistics, Computer Science, or Math. You should also possess at least 2 years of industry experience in developing machine learning models, experience with frameworks like PyTorch or TensorFlow, and a solid understanding of production-level machine learning.

DoorDash promotes a collaborative and inclusive work culture for Software Engineers, Machine Learning - Fraud. You can expect to work in a team-oriented environment where diversity of thought is highly valued. With a focus on innovation, team members are encouraged to share unique perspectives that drive impactful solutions, all while enjoying comprehensive benefits and support.

In the Software Engineer, Machine Learning - Fraud role at DoorDash, you will work on projects that involve developing production machine learning solutions to enhance the safety of our logistics operations and improve customer experiences. This includes tackling complex challenges in the fraud detection domain, leveraging both classical and deep learning methods.

At DoorDash, professional development for a Software Engineer, Machine Learning - Fraud is a priority. Employees have access to mentorship opportunities, training resources, and collaboration with experts in machine learning. Additionally, the company encourages team members to take part in projects that align with their interests and career goals, helping to foster growth and innovation.

When answering this question, focus on discussing the problem you aimed to solve, the data you used, algorithms implemented, and the impact your solution had. Highlight aspects that relate specifically to fraud detection or your machine learning workflow.

Discuss your methods for feature selection, such as using domain knowledge, statistical tests, and utilizing techniques like Lasso or decision trees. Relate your answer to how these methods could specifically apply to detecting fraud in transactions.

Share examples of challenges like data drift, model performance over time, or integration issues. Elaborate on how you addressed these challenges, especially in high-stakes environments like fraud prevention.

Explain your validation strategies like cross-validation, A/B testing, or developing metrics specific to fraud detection, such as precision and recall. Provide insight into why validation is crucial for maintaining model accuracy in real-world applications.

Discuss specific projects where you utilized either TensorFlow or PyTorch. Highlight your familiarity with building and training deep learning models and how you applied these frameworks to real-world problems, particularly in relation to fraud detection.

When addressing imbalanced datasets, be sure to mention techniques like undersampling, oversampling, and using algorithms that are sensitive to class distribution. Relate this back to fraud detection, where fraudulent cases are often rare compared to legitimate transactions.

Talk about your methods for keeping up with the latest research, including following ML blogs, attending conferences, and participating in online courses. Share how these efforts help you bring fresh insights to your work at DoorDash.

Detail your experience with any JVM languages, including projects completed using those languages. Highlight how they may enhance your contributions as a software engineer at DoorDash, particularly in integrating ML solutions.

Explain your understanding of algorithmic bias and strategies you might employ to mitigate it, such as thorough testing and regular audits of model outputs. Discuss how this is particularly relevant in creating fair and ethical fraud detection systems.

Express your enthusiasm for DoorDash’s mission and how your skills align with the company’s goals in the fraud detection space. Discuss your passion for working in innovative and fast-paced environments like the one at DoorDash.