Data Scientist – Forecasting

GSPANN is hiring a Data Scientist with expertise in Forecasting to design, build, and deploy production forecasting models across the full Machine Learning (ML) lifecycle. The role focuses on hands-on time-series and hierarchical forecasting, MLOps standardization, and feature-store design for exogenous signals like price, promotion, weather, and events.

• Collaborate with commercial Subject Matter Experts (SMEs) across Sales, Supply, Finance, and Revenue Management to frame forecasting problems, translate questions into testable hypotheses, and define success criteria and Key Performance Indicators (KPIs).
• Design, build, validate, and deploy forecasting and ML models, owning the full lifecycle from problem definition and feature engineering to Machine Learning Operations (MLOps), monitoring, and continuous improvement.
• Drive forecasting at scale by standardizing data pipelines, model templates, and evaluation frameworks so solutions roll out across regions and markets with minimal rework.
• Ensure responsible Artificial Intelligence (AI) standards across the Software Development Life Cycle (SDLC), covering data quality, experimentation, explainability, bias checks, governance, and documentation.
• Develop performant, reliable, and cost-efficient pipelines (batch and near-real-time) and model serving in partnership with platform and engineering teams.
• Manage vendor and strategic partner relationships to influence roadmaps and ensure solutions meet long-term needs.

• 5–9 years delivering end-to-end analytics and data science on large datasets, including production deployment.
• Apply strong Python skills (pandas, NumPy, scikit-learn); experience with PySpark, Databricks, and the Microsoft Azure ecosystem (ML, Azure Data Factory (ADF), Functions, DevOps) is a strong plus.
• Demonstrate solid MLOps experience across experiment tracking, model registry, CI/CD, automated testing, data and concept drift monitoring, alerting, and rollback.
• Work with strong SQL and data pipeline skills, including familiarity with medallion and lakehouse patterns and performance/cost optimization.
• Develop hands-on time-series and hierarchical forecasting, including seasonal linear regression integrated with weather effects, promotional uplift, and event-based adjustments.
• Utilize promotion and price analytics by building features for uplift, price elasticity, and seasonality/holiday impacts, and modeling new product introductions (NPI), discontinuations, and intermittent demand patterns.
• Manage feature-store design for exogenous signals (price, promo, weather, events, macro indices, media) to standardize reuse across models and markets.
• Apply multi-level evaluation from SKU/store/day to channel/region, with rigorous backtesting, probabilistic metrics, and FVA; demonstrate fluency with WAPE, MAE, sMAPE, MASE, prediction intervals, and calibration.
• Demonstrate forecast operations in production — champion-challenger testing, planner-facing explainability, exception management, and governed human-in-the-loop overrides.