Data Scientist

Position: (USA)Staff, Data Scientist
Position Summary...

What you'll do...

About the Role We're looking for a Staff Data Scientist to design and build advanced forecasting models to ensure accurate financial planning and analysis (FP&A) that power critical business decisions. You'll build and deploy state-of-the-art time series models (classical + ML + deep learning), drive explainability and trust (XAI), and explore next-generation approaches such as graph neural networks for spatiotemporal and relational forecasting problems.

You'll partner closely with engineering, product, and stakeholders to deliver measurable impact t You'll Do

* Design and deploy statistically and ML models to address high-impact financial forecasting needs, ensuring alignment with Walmart's business objectives.

* Perform statistical analysis across large data sets and within defined segments to empower data driven decisions.

* Own E2E forecasting lifecycle, including scoping, feature engineering, model development, experimentation, monitoring and ongoing performance optimizations.

* Develop advanced time series solutions using:

* Statistical methods (ETS, ARIMA/SARIMA, State Space Models)

* ML approaches (GBMs, Random Forests, linear/elastic models with engineered time features)

* Deep learning (RNN/LSTM/GRU, Temporal Convolutional Networks (TCNs), TimesFM)

* Probabilistic forecasting and uncertainty quantification (quantile regression, Bayesian approaches, conformal prediction, prediction intervals)

* Build explainable forecasting systems: model interpretability, feature attribution, drivers of change, scenario analysis, and stakeholder-facing narratives.

* Apply graph-based and spatiotemporal modeling where relationships matter: GNNs, temporal graphs, graph embeddings.

* Establish strong evaluation and monitoring: backtesting, leakage prevention, stability checks, drift detection, calibration of uncertainty, and post-deployment performance tracking.

* Drive best practices in MLOps and production readiness: reproducible pipelines, scalable training/inference, model versioning, and governance.

* Build Agentic workflows to enable chat based forecasting explainability and scenario planning.

* Collaborate with cross-functional partners including Product, Business, Data Science and Engineering.

* Mentor other data scientists, set modeling standards, and influence technical direction across teams.

What You'll Bring (Required)

* 8+ years in data science / applied ML (or PhD + 5 years), with deep hands-on exposure to forecasting and predictive modeling.

* Demonstrated experience delivering production grade ML models with measurable business outcomes.

* Strong knowledge of time series topics: seasonality, hierarchies, intermittent demand, holidays/events, promotions, missingness, outliers, anomaly detection, and regime changes.

* Hands-on experience with deep learning frameworks (PyTorch or Tensor Flow) and modern architectures for time series.

* Practical experience with explainable AI methods and communicating model reasoning to non-technical stakeholders.

* Excellent coding skills in Python; strong grasp of software engineering fundamentals (testing, packaging, code reviews).

* Ability to translate ambiguous business problems into rigorous modeling plans and deliver results.

* High attention to detail and an ownership mindset in managing multiple high-impact projects.

Preferred Qualifications

* Experience with graph neural networks (PyG/DGL), spatiotemporal GNNs, or temporal graph learning.

* Experience with causal inference or decision-focused forecasting (uplift, impact estimation, counterfactuals, policy evaluation).

* Familiarity with large-scale data/compute:
Spark, distributed training, feature stores, GPU workflows.

* Experience building human-centered explainability: dashboards, driver decomposition, "why changed" analysis, model cards.

* Publications, patents, or open-source contributions in time series, XAI, or graph learning.

Key Skills / Tech Stack

* Proficiency in Python, Sql and data visualization tools.

* Experience using PyTorch/Tensor Flow; scikit-learn; XGBoost/LightGBM and other models for production grade models.

* Experience building solutions with time series libraries…