Sr. Data Scientist

About the role

We're looking for a hands-on Data Scientist to own and advance the demand-forecasting and inventory-optimization models behind a live inventory-intelligence platform used daily by a retail client. The platform is in production; your mission is to keep making it sharper — improving forecast accuracy against real weekly sales data, and translating business logic into models the client understands and trusts. This sits at the intersection of applied ML, MLOps, and supply-chain domain.

What you'll do

Own the demand-forecasting models (weekly operational + monthly planning horizons, segmented by demand pattern) and drive measurable accuracy gains.

Interrogate the data.

Analyze incoming weekly and monthly sales data to validate and refine the business logic against real-world demand patterns.

Model hard demand — intermittent, lumpy, and seasonal SKUs; handle out-of-stock / censored demand honestly rather than papering over it.

Improve the inventory logic — safety-stock sizing, reorder points, and service-level trade-offs that map to real purchasing decisions.

Keep the pipeline healthy — automated retraining, experiment tracking, model-promotion gates, leakage guards, and accuracy monitoring.

Report accuracy honestly and explain model behavior to non-technical stakeholders in plain business language.

Partner with engineering to ship model outputs into the client-facing dashboard.

What you bring

3+ years of applied data science / ML, with models that have run in production (not just notebooks).

Strong Python and practical ML — gradient boosting (e.g. LightGBM/XGBoost) and time-series forecasting.

Solid statistics and forecasting fundamentals — seasonality, intermittent demand, error metrics (MAPE / wMAPE / bias), and calibration. Comfort with SQL and large tabular data (parquet, cloud storage). An MLOps sensibility — reproducible pipelines, experiment tracking, monitoring, and a healthy paranoia about data leakage.

Bonus points

Retail, e-commerce, supply-chain, or demand-planning domain experience. Amazon Selling Partner API / marketplace data; Inventory Planner or similar. Orchestration (Dagster / Airflow), MLflow, AWS (S3), and awareness of modern dashboards (Next.js). Inventory theory — safety stock, service levels, reorder-point / EOQ thinking

Work Location: In person