ironquill.tech/board

$ cat jobs/staff-data-scientist-csc-generation-0874133fc97b.json

Staff Data Scientist

csc generation·Canada·Canada, United States·senior
Apply on himalayas → Get AI match score →
CSC Generation is the AI-native holding company re-engineering omnichannel retail. We acquire iconic brands and transform them with Genesis, our operating platform combining a Data Fabric, Automation Engine, proprietary tools, and shared services to modernize operations, elevate customer experience, and expand margins. With $1B+ in revenue across 13 brands, our portfolio includes Sur La Table, Backcountry, One Kings Lane, and others that serve as real-world innovation labs. Reports to : Director of Finance and Business Intelligence Location: Remote — US or Canada About the Role As our Staff Data Scientist , you will design and ship production pricing systems such as demand forecasting, price elasticity modeling, dynamic pricing and the experimentation infrastructure needed to measure whether they actually work. This is a hard, high-stakes problem: your models will directly influence margin and revenue decisions across a portfolio of brands operating at scale. You will own the full arc from framing ambiguous business problems as well-defined ML tasks through to monitoring models that hold up in production. At six months, success looks like at least one pricing model shipped to production with measurable business impact and an experimentation framework in place that your stakeholders trust. If you have spent time building pricing systems from the ground up, not just consuming them, and you care deeply about rigorous causal inference and honest model evaluation, this role was written for you. What You'll Do Design and build production ML systems for pricing, demand forecasting, and related revenue problems Frame ambiguous business problems as well-defined ML tasks with clear success criteria and measurable outcomes Set the standard for model evaluation, validation, and monitoring — including knowing when CV metrics are misleading and when holdout testing is the only honest answer Build robust predictive models across classification, regression, time series, and causal