Data Scientist III

PODS•Clearwater, FL

About The Position

PODS is building the analytical infrastructure to understand customer behavior, quantify price elasticity, and inform daily commercial decisions across our long-distance and local moving businesses. As a Data Scientist 3, you are a senior individual contributor on the Revenue Science team, reporting to the Director of Pricing Strategy and Analytics. You have built and deployed real models end to end and can operate without mature infrastructure in place — designing the pipeline, the model, and the measurement, shipping them to production, and owning the stakeholder relationship. You’ll set the technical standard for the team, mentor earlier-career data scientists, and take on the hardest modeling, optimization, and measurement problems behind pricing decisions worth millions of dollars to the business.

Requirements

Expert Python and advanced SQL: Expert-level Python (pandas/Polars, NumPy, scikit-learn, statsmodels) and advanced SQL, including performance work on very large tables.
Statistical modeling depth: Strong foundation in generalized linear models, hierarchical models, and forecasting, with the judgment to defend a specification — not just fit one.
Causal inference and experiment design: Deep experience with holdouts, geo/cluster randomization, power analysis, difference-in-differences, and similar quasi-experimental methods.
Optimization and operations research: Working proficiency with LP, QP, and MIP — able to formulate a pricing or capacity decision as an optimization problem and solve it with Gurobi, CVXPY, OR-Tools, or similar.
Production deployment: Git, CI, orchestration (Airflow, Databricks, or similar), containers, and model monitoring with drift detection.
Distributed compute and cloud depth: PySpark/Databricks or equivalent at scale, with depth on a modern cloud data platform; our stack is Snowflake and Azure/Databricks, and deep AWS or GCP experience transfers fine.
AI-accelerated analytical workflows: Demonstrated use of AI tools (Claude, Cursor, Copilot, or similar) to accelerate code, query, and documentation work, with the judgment to set team standards for when AI output requires verification.
Executive communication: Ability to carry a recommendation from analysis to decision with senior stakeholders — plain language, quantified uncertainty, and a defensible answer under pressure.
Master’s or PhD in a quantitative field (Computer Science, Data Science, Operations Research, Statistics, Econometrics, Industrial Engineering, Economics, or similar), or equivalent applied experience.
7+ years of post-academic experience building and deploying models end to end as a senior individual contributor.
Has owned something end to end — pipeline, model, deployment, and the stakeholder relationship — not just the modeling slice.
Has built where the data infrastructure was not ready and shipped anyway, and can quantify the business impact of their own work and explain how it was measured.
Comfortable with ambiguity and moving before all the information is in.

Nice To Haves

A pricing background is a plus, not a requirement.
Experience in moving, logistics, e-commerce, travel/hospitality, or other capacity-constrained consumer businesses is also a plus.

Responsibilities

Build, deploy, and monitor the team’s core models — elasticity, demand, conversion, and forecasting — owning the pipeline, the model, the deployment, and the stakeholder relationship.
Formulate and solve optimization problems (linear, quadratic, and mixed-integer programming) for pricing and capacity decisions using tools such as Gurobi, CVXPY, or OR-Tools.
Establish model monitoring and drift detection so deployed models stay trustworthy, and rebuild or retire them when they do not.
Define how experiments are designed and analyzed across the team: holdouts, geo/cluster randomization, power analysis, and metric definitions.
Choose and defend identification strategies (difference-in-differences and similar quasi-experimental methods) when randomization is not feasible.
Arbitrate methodological questions on high-stakes measurement, and make the call when evidence is incomplete and a decision cannot wait.
Design and ship production-grade pipelines and data models — git, CI, orchestration (Airflow, Databricks, or similar), and containers — without waiting for mature infrastructure.
Scale analytical work with distributed compute (PySpark/Databricks or equivalent) and performance-tune SQL on very large tables.
Build the reusable assets — feature tables, model libraries, evaluation harnesses — that make the rest of the team faster.
Own senior stakeholder relationships: present recommendations to commercial leadership, quantify the business impact of shipped work, and explain how it was measured.
Mentor earlier-career data scientists on methods, code, and judgment, and review the team’s highest-stakes analyses before they ship.
Scope ambiguous commercial questions into tractable analytical plans, moving before all the information is in.