Machine Learning Engineer

Stellantis is hiring a Machine Learning Engineer / Data Scientist for an onsite role in Auburn Hills, MI. In this position, you will build statistical models and simulations to support Vehicle Configuration Optimization (VCO) using a customer-level preference simulation engine, with the goal of producing optimized Vehicle Order Guides (VOGs) for upcoming model years.

What you’ll do

• Build and run large-scale simulations, including scenarios such as 50,000 synthetic customers, to model vehicle purchase behavior
• Develop statistical and machine learning models using Databricks
• Perform exploratory data analysis and feature engineering on complex datasets
• Leverage datasets that include historical vehicle sales, competitive sales data, feature-level willingness-to-pay data, and customer preference models
• Translate model outputs into optimized Vehicle Order Guides (VOGs) that inform product configuration decisions
• Collaborate closely with Data Engineering to refine and leverage curated datasets
• Communicate insights and model recommendations to business stakeholders
• Continuously evaluate and improve model accuracy and assumptions

Required qualifications

• Bachelor’s Degree required
• Minimum 5 years of experience in data science, machine learning, or applied statistics
• Strong experience with Databricks (critical requirement)
• Proficiency in Python including Pandas, NumPy, scikit-learn, and PySpark
• Strong SQL skills
• Solid background in statistical modeling, simulation techniques, and experimental design
• Ability to translate analytical results into business decisions

Technology stack

• Databricks
• Python, including Pandas, NumPy, scikit-learn, and PySpark
• SQL
• Spark (via PySpark)

Preferred qualifications

• Experience with choice modeling, conjoint analysis, or demand modeling
• Background in automotive, pricing, or product optimization analytics
• Experience working with large-scale simulation frameworks
• Familiarity with Spark and distributed computing
• Exposure to MLOps or model productionization