Full-Stack Data Scientist, Hardware Reliability (Starlink)
Core
Build data pipelines, models, and production scoring systems to improve the reliability of Starlink customer hardware (dishes, routers, power supplies) using telemetry, manufacturing, and unstructured support data.
Role type
Full-stack data scientist (hardware reliability)
Builds
Production scoring systems, LLM-backed tools (retrieval, tool calling, agentic workflows), and internal analytics APIs for hardware reliability teams.
Domain
Satellite broadband (Starlink) / Hardware reliability / Aerospace
Deliverable
production ML models
Required skills
Python, SQL, production data pipelines, machine learning model deployment, software engineering fundamentals, relational databases, data pipeline orchestration, large-scale data processing, LLM production (RAG, tool calling, agentic workflows), unstructured data extraction, statistical analysis, reliability methods (survival analysis, calibration, experimental design)
Preferred skills
Go, Java, C++, C#, advanced software system design, high-performance code optimization
Technologies
Python, SQL, LLM frameworks, data pipeline orchestration tools, internal codebases
Responsibilities
Build and maintain data pipelines and infrastructure for hardware reliability; design and operate feature, label, and scoring data; train and iterate models for hardware failure prediction; deploy and monitor models in production; contribute to core software repositories; build internal analytics tools and APIs; collaborate across engineering, production, and supply chain teams to implement reliability improvements.
Seniority
Mid-Senior, hands-on IC