Full-Stack Data Scientist, Hardware Reliability (Starlink)

RESPONSIBILITIES:

• Build and maintain data pipelines, infrastructure, and software to objectively increase the reliability of Starlink customer hardware
• Design and operate the feature, label, and scoring data that reliability models train and serve on, including backfills, late-arriving data, and data quality
• Train, validate, and iterate models for hardware failure prediction both in production line and field
• Deploy and operate models in production, including scheduled scoring, versioning, monitoring, and iteration based on live results
• Contribute to critical Starlink software repositories, working in core internal and customer-facing codebases
• Build internal analytics tools, APIs, and applications used by hardware, production, quality, supply chain, and customer experience teams
• Collaborate across engineering, production, test, inventory, quality, supply chain, and customer experience to turn analyses and predictions into design, process, and support changes
• Identify customer needs, gather requirements, and execute independently in a dynamic environment with changing constraints and tight deadlines
• BASIC QUALIFICATIONS:
• Bachelor’s degree in computer science, data science, mathematics, electrical engineering, or other STEM discipline
• 4+ years of professional experience in data science, machine learning, or software engineering
• Strong proficiency in Python and SQL
• Experience building production data pipelines
• Experience deploying machine learning models in production
• PREFERRED SKILLS AND EXPERIENCE:
• Master’s degree in computer science, data science, mathematics, or other STEM discipline (or PhD with 3+ years of relevant experience)
• Strong understanding of software engineering fundamentals including data structures, algorithms, system design, concurrency, and writing high-performance, maintainable code
• Experience with relational databases, data pipeline orchestration, and large-scale data processing
• Experience deploying, versioning, and monitoring machine learning models in production, including scheduled scoring jobs whose outputs are consumed by other systems