New opportunity

Data Scientist

Skechers · Manhattan Beach, United States of America

About this role

The Data Scientist joins Skechers’ Data and Analytics team, reporting to the Sr. Director, BI and Reporting. The role combines statistical rigor and an engineering-first approach to turn business challenges into production-ready machine learning solutions and advanced analytics for core data products.

Responsibilities

Develop, test, document and improve data science applications and models across the lifecycle, from exploration and experimentation through deployment and operation. Write clean, scalable production code and collaborate with data engineers on pipelines and model deployment.
Use large language models and retrieval-augmented generation to extend data products and automate internal workflows. Monitor production models for drift and bias to maintain accuracy, performance and ethical use.
Lead experimental design, A/B testing and causal inference work to assess product changes and model impact. Conduct ad hoc analysis to identify patterns, anomalies, trends and business opportunities.
Build and test scalable BI reports and analytics solutions, prototype in response to user feedback, document technical and business outcomes, and explain findings to nontechnical leaders. Work with data and BI engineers, data product managers, business leads, data stewards and users.

Required qualifications

Bachelor’s, master’s or Ph.D. in a quantitative field such as computer science, statistics, mathematics or physics, or equivalent industry experience; 3+ years in a professional data science role, including moving models from research into production.
Proficiency in Python, Spark, NumPy, Pandas, scikit-learn, XGBoost, TensorFlow and PyTorch; grounding in statistics, probability, optimization, linear algebra, classification, regression, clustering and multivariate methods. Expertise in relational data modeling and advanced SQL.
Experience with Agile/SCRUM, DSML automation and MLOps tools; familiarity with retail, supply chain, digital and financial analysis, predictive analytics and BI visualization. Experience with Snowflake, BigQuery, AWS SageMaker, Hugging Face, LangChain, LlamaIndex, OpenAI/Gemini APIs and cloud-based analytics technologies. Strong independent analysis, communication, presentation, organization and collaboration skills are expected.

Preferred

5+ years in a professional data science role involving advanced analytics applications. The listed location is Manhattan Beach, CA; the stated pay range is $95.000 – $130,000/year USD.

For full details about this role, click “Apply on company site” to view the employer’s original posting.

Skills for this role

PythonSparkNumPyPandasscikit-learnXGBoostTensorFlowPyTorchSQLRelational data modelingStatisticsProbabilityOptimization algorithmsLinear algebraClassificationRegressionClusteringMultivariate analysisMachine learningLarge language modelsRetrieval-augmented generationExperimental designA/B testingCausal inferenceModel deploymentModel drift monitoringBias monitoringData pipelinesPredictive analyticsBusiness intelligence visualization tools (BI) visualization tools (BI)

Checking your profile…

Preparation suggestions, not the employer’s actual interview questions. Always check the original posting for current requirements.

Understand the role

What matters most

Move ML solutions into production
The role spans ideation, analysis, model creation, deployment and operation. It calls for maintainable, scalable code and collaboration with data engineers on pipelines and deployment.
Extend data products with LLMs and RAG
The posting explicitly calls for using LLMs and retrieval-augmented generation to broaden data products and automate internal workflows, while monitoring production models for drift and bias.
Measure impact rigorously
Leading experimental design, A/B testing and causal inference work is a stated duty, with the aim of measuring the impact of product changes and model performance.
Deliver tested analytics and BI
The role includes ad hoc analysis, rapid prototypes shaped by user feedback, and scalable enterprise reports and analytics solutions tested before release.
Turn findings into business decisions
The posting emphasizes documenting objectives and outcomes, explaining model results to nontechnical leaders, and working with engineers, product managers, business leads and users.

Likely challenges to think through

Speed versus production reliability
Possible challenge, not a reported Skechers problem: quick prototypes may need substantial work before they meet the posting’s testing, documentation and scalability expectations. Prepare to explain how you would decide what to prototype, then validate and hand off for production.
Attributing impact convincingly
Possible challenge, not a confirmed obstacle: product changes and model effects can be hard to separate from other influences. Prepare to explain how you would choose between an A/B test and a causal-inference approach, check assumptions, and communicate uncertainty.
Keeping AI useful over time
Possible challenge, not a reported issue: extending products with LLMs and RAG must be balanced against answer quality and the stated duty to monitor drift and bias. Prepare an evaluation and monitoring approach, including when findings would prompt a change.
Serving different audiences without losing rigor
Possible challenge, not a confirmed employer problem: engineers need implementation detail while business leaders need actionable conclusions. Prepare to describe how the same analysis would be documented technically and presented with clear implications and limitations.

Practice your answers

Technical practice

Suggested practice: How would you frame a business problem as a machine learning task?

What to coverThe role calls for turning business challenges into production-ready ML solutions. Discuss how you would define the decision, select an appropriate classification, regression or clustering approach, and decide whether a simpler analytic method would suffice.

Suggested practice: How would you evaluate whether a model is ready to deploy?

What to coverThe posting emphasizes statistical rigor and the full model lifecycle. Explain your choice of validation design, baseline and performance measures, including how you would check for leakage and whether results generalize.

Suggested practice: How would you design and optimize relational data for an analytics use case?

What to coverAdvanced SQL and relational data modeling are stated requirements. Discuss table grain, joins, data-quality checks and how you would diagnose and improve a slow analytical query.

Suggested practice: How would you prepare large-scale data for model development?

What to coverThe role uses Python and Spark and involves collaboration with data engineers on pipelines. Describe how you would make transformations reproducible, check inputs and outputs, and choose between local and distributed processing.

Suggested practice: How would you move an experiment into a maintainable production model?

What to coverThe posting calls for production-grade code, deployment collaboration and experience with MLOps tools. Cover testing, documentation, versioning, deployment handoffs and operational considerations without assuming a particular platform is used for every model.

Suggested practice: How would you assess an LLM and retrieval-augmented generation solution?

What to coverLLMs and RAG are named as ways to extend data products and automate internal workflows. Explain when retrieval is useful, how you would evaluate answer quality and source grounding, and what trade-offs you would examine before deployment.

Suggested practice: How would you detect and respond to model drift or bias?

What to coverProduction monitoring for accuracy, performance and ethical use is an explicit duty. Discuss what you would monitor, how you would investigate a concerning change, and when you would recommend remediation or retraining.

Suggested practice: How would you design an experiment to measure a product change?

What to coverThe role leads experimental design, A/B testing and causal inference work. Address the outcome measure, assignment or comparison strategy, statistical uncertainty and threats to a causal interpretation.

Suggested practice: How would you investigate an unexpected business trend?

What to coverThe posting includes independent ad hoc analysis of patterns, relationships and anomalies across data formats and platforms. Walk through data checks, alternative explanations and how you would distinguish a useful finding from noise.

Suggested practice: How would you validate an enterprise BI report before release?

What to coverBuilding scalable analytics solutions and thoroughly testing reports are stated duties. Explain how you would verify metric definitions and results against source data, test user-facing behavior and document the report’s business meaning.

Behavioural practice

Suggested practice: Tell me about aligning stakeholders on an ambiguous analytics problem.

What to coverThe role works with engineers, data product managers, business leads, stewards and users. Use a real situation, your actions to clarify the decision and reconcile differing needs, and the resulting agreement or outcome.

Suggested practice: Tell me about explaining a technical finding to nontechnical leaders.

What to coverThe posting emphasizes storytelling and translating model outputs into actionable strategies. Describe the situation, how you presented uncertainty and implications without obscuring the evidence, and what resulted.

Suggested practice: Tell me about changing a prototype in response to user feedback.

What to coverQuick prototyping and agile adjustment to user requests are stated duties. Give a real example of the initial need, the feedback and actions you took, and the result for users.

Suggested practice: Tell me about prioritizing competing work with limited oversight.

What to coverThe role expects independent analysis, organization and timely, accurate completion. Explain the situation, how you weighed urgency against quality and communicated trade-offs, and the outcome.

Suggested practice: Tell me about handling a problem after an analytics solution went live.

What to coverThe posting calls for model operation, monitoring, report testing and change management. Use a real situation to describe how you investigated the issue, coordinated a response and documented the result or lesson.

Sharpen your knowledge

Topics to prepare

Statistical and machine-learning foundations
Review probability, optimization, linear algebra, and when to use classification, regression, clustering or multivariate methods. As a revision exercise, compare two methods for a hypothetical predictive-analytics question and explain the assumptions and evaluation measures. The posting explicitly calls for these foundations and techniques.
Experimental design and causal inference
Sketch an A/B test for a hypothetical product change: define the outcome, assignment, possible confounders and how you would interpret the result. This prepares for the stated responsibility to lead experiments and assess the impact of changes and models.
Production model lifecycle
Map a model from exploration through deployment and operation, then specify what drift and bias signals would prompt investigation. The role explicitly spans the full lifecycle and requires monitoring production models.
LLMs and retrieval-augmented generation
Diagram how retrieval, an LLM and checks on the resulting answer could support a hypothetical internal workflow. This is a preparation exercise, not a specified Skechers project; the posting explicitly calls for LLMs and RAG in data products and workflow automation.
Business-facing analytics
For a hypothetical retail or supply-chain question, define a useful measure, a likely data limitation and a decision-oriented explanation of the result. The posting calls for familiarity with those business areas, BI reporting and communication to nontechnical leaders.

Skills to revise

Python and Spark analysis
Practice cleaning and exploring a sample dataset in Python using NumPy or Pandas, then express a larger-scale transformation in Spark. These are explicitly named tools for programming and analysis in the posting.
Model implementation and testing
Build a small classification or regression example with scikit-learn or XGBoost; document the evaluation, test the code and explain what would need to change for production. The posting names these frameworks and emphasizes maintainable code and models that move beyond research.
Advanced SQL and relational modeling
Design related tables for a hypothetical analytics question, write a multi-table aggregation and inspect how its joins or filters could affect performance. The posting explicitly asks for relational data-modeling expertise and advanced SQL optimization.
Deployment and monitoring workflow
Draft a handoff for data engineers covering pipeline inputs, model outputs, deployment checks and drift or bias monitoring. As a practical exercise, this reflects the stated collaboration on pipelines and deployment and the requested MLOps experience; it does not assume a particular Skechers platform.
BI reporting and presentation
Prototype a report from sample results, check its figures against the underlying data, and give a brief explanation for a nontechnical leader. The role explicitly requires tested BI solutions, rapid feedback-driven prototyping and clear communication of actionable findings.

Your preparation checklist

Map the role requirements to evidence
Prepare a one-page list of examples you can substantiate for production data science, Python and Spark, advanced SQL, statistical methods, MLOps, and collaboration. Note gaps honestly; the posting requires 3+ years of professional data science experience moving models to production.
Build a production-model walkthrough
Select a real project and outline the business problem, data, model choices, evaluation, testing, deployment, and results. Be ready to explain code maintainability, pipeline collaboration, and how you monitored or would monitor drift and bias.
Practice a data and SQL design exercise
Sketch a relational model for a hypothetical retail analytics question, then write sample SQL for joins, aggregations, and performance optimization. This is preparation, not a stated Skechers interview exercise.
Draft an experiment plan
Choose a hypothetical product or model change and produce a short A/B-test plan covering the outcome metric, assignment, analysis, and threats to causal interpretation. The role explicitly leads experimental design and impact measurement.
Sketch an LLM and RAG solution
Prepare a one-page design for a hypothetical internal workflow: retrieval sources, evaluation approach, safeguards, and operational monitoring. Label it as a proposal rather than an existing Skechers project.
Prepare an analytics-to-decision example
Bring an example of ad hoc analysis or a BI report, showing data checks, a finding, user feedback, and a concise recommendation for a nontechnical leader. The posting emphasizes tested reports, rapid prototyping, and clear documentation.
Rehearse cross-functional communication
Practice explaining one technical trade-off and one project outcome at two levels: for data or BI engineers and for business leads. Include how you prioritized work and incorporated feedback in an Agile setting.
Write targeted questions for the employer
Ask which core data products need support, how model deployment and monitoring are handled, how experiment results inform decisions, and how data scientists work with engineers, product managers, and business users.

Availability last checked: Nov 4, 2026. Status can change after a check; confirm on the employer’s website. Report an issue · Our checking policy

KEEP EXPLORING

Similar AI jobs

Related skills and specializations in United States of America. Matched to this role, not your profile.

Explore more jobs

Finding similar opportunities…