This description is a summary of our understanding of the job description. Click on 'Apply' button to find out more.

Role Description

Stord is building ML capabilities that directly power our cloud-based supply chain platform, which handles over $10B in commerce annually. You'll work alongside a Machine Learning Engineer to own the full data science lifecycle — from exploratory analysis and model development through production deployment and ongoing performance improvement. This is a high-impact, hands-on role on a small team where your work will directly shape how millions of shipments are planned, routed, and fulfilled.

As a Senior Data Scientist at Stord, you will own the scientific rigor behind our ML features — designing experiments, developing models, and translating complex data findings into actionable product decisions. You'll work closely with an ML Engineer who owns productionization, but you'll remain deeply involved through deployment, monitoring, and iteration. This is a rare opportunity to apply data science directly to hard logistics problems with immediate, measurable customer impact.

What You'll Do

• Data Analysis & Problem Framing
  • Conduct exploratory data analysis to validate assumptions, surface insights, and identify data quality issues before they affect model development
  • Answer specific business questions with rigorous, data-driven analysis
  • Define success metrics and evaluation frameworks in collaboration with product and engineering stakeholders
  • Translate ambiguous business problems into well-scoped data science problems
• Model Development
  • Design, build, and evaluate predictive models for logistics use cases: delivery time estimation, demand forecasting, routing optimization, capacity planning
  • Own model quality — feature selection, validation methodology, bias detection, and performance benchmarking
  • Run structured experiments to validate improvements before production promotion
  • Contribute to improving existing production models using performance data and operational feedback
• Production Involvement
  • Write production-quality Python code, not just notebook prototypes
  • Partner closely with the ML Engineer through deployment — your involvement doesn't end at handoff
  • Monitor model performance in production and own the scientific response to drift or degradation
  • Contribute to A/B testing design and interpretation of results
• Technical Translation & Collaboration
  • Communicate model behavior, limitations, and tradeoffs clearly to engineers, product managers, and business stakeholders
  • Serve as Stord's subject matter expert in data science and ML — you are expected to lead this domain, not just contribute to it
  • Present findings, recommendations, and model decisions to leadership and executives, translating technical complexity into business impact and strategic context
  • Document technical decisions in ways accessible to non-data scientists
  • Participate in sprint planning, code review, and architectural decisions for AI/ML features
  • Help other engineers build intuition around statistical methods and ML approaches

Qualifications

• 5+ years of data science experience with models shipped to and maintained in production
• Expert-level Python — production code, not just analysis scripts
• Strong SQL — complex queries, performance optimization, BigQuery and Postgres experience
• Deep understanding of statistical fundamentals and ML model evaluation
• Experience with cloud ML platforms, preferably GCP
• Familiarity with logistics, e-commerce, fulfillment, or supply chain domains — you understand what on-time delivery, carrier performance, and demand variability actually mean operationally

Requirements

• Experience working embedded with software engineering teams rather than in a traditional data science org
• Familiarity with feature engineering for real-time inference
• Elixir or TypeScript exposure, or comfort operating in polyglot engineering environments
• CI/CD and DevOps familiarity
• Experience with monitoring tools (Datadog or equivalent) applied to model performance
• Contributions to model improvement, not just greenfield development