Associate Data Scientist

VF Corporation is looking for an Associate Data Scientist to join our team in Stabio, Switzerland.

Let’s talk about the role!

As an Associate Data Scientist at VF Corporation, you will focus on demand forecasting for our fashion and retail business: anticipating what consumers will buy, where and when, across brands, channels and markets, including products and stores with little or no sales history.

In this role you will:

• Demand forecasting: build forecasting models at product, size, store, channel and market level across Wholesale and Direct-to-Consumer, supporting buying, allocation, replenishment and planning.
• Cold-start forecasting: contribute to the team's work forecasting demand for new products, collections, stores and markets with no sales history, a core challenge in fashion and central to this role.
• State-of-the-art methods: benchmark and product ionize zero-shot and foundation models for time series against classical and machine learning baselines.
• Collaboration:
  
  contribute to AI and Generative AI use cases, help monitor deployed models, and work with Data Engineers, Dev Ops, Product, Planning and Merchandising to drive adoption, presenting results and their limitations clearly to stakeholders.

This is an associate-level position: you will work within an established forecasting team, with a senior data scientist as your mentor, growing toward end-to-end ownership of your models.

(we don't expect all of these; the more you bring, the better):

• Demand Forecasting in Fashion and Retail: hands-on experience with short product life cycles, strong seasonality, promotions and markdowns, size and colour curves, sparse demand and hierarchical forecasting.
• Cold-Start Forecasting: attribute- and similarity-based approaches, product embeddings, like-for-like matching, transfer learning and pre-season planning.
• Zero-Shot and Foundation Models for Time Series: up-to-date knowledge of the state of the art (e.g., TimeGPT, Chronos, TimesFM), and the ability to benchmark zero-shot versus fine-tuned performance against strong baselines.
• Time Series Foundations: statistical (ARIMA, ETS, Croston), machine learning and deep learning (e.g., Gradient Boosting, DeepAR, N-BEATS, N-HiTS) approaches, plus probabilistic forecasting, backtesting and metrics such as WMAPE and MASE.
• Python and ML Libraries: strong Python with pandas, Num Py, scikit-learn, PyTorch or Tensor Flow, and forecasting libraries such as stats models, Stats Forecast, Neural Forecast.
• Statistics and SQL: core ML concepts, hypothesis and A/B testing, hyperparameter optimisation, and efficient SQL on large-scale structured data.
• Generative AI basics: how LLMs are applied in practice (prompting, embeddings, RAG).

(again, not a checklist; the more you bring, the better):

• Databricks Lakehouse Platform: our core platform. Hands-on experience product ionizing data and machine learning solutions with Python, PySpark, SQL, Delta Lake, notebooks and Workflows.
• Forecasting at Scale: training and scoring large numbers of time series in parallel on Spark, with Databricks Feature Engineering, MLflow and Unity Catalog.
• Pipelines and MLOps: scalable data and feature pipelines with Databricks Jobs and Pipelines, plus Git and CI/CD in Azure Dev Ops.
• Rapid Prototyping: lightweight front-ends (Streamlit + Databricks Apps or equivalent) so stakeholders can interact with and validate forecasts.

• Business Understanding and Impact: translate planning, buying, allocation and merchandising problems into data science solutions with measurable value.
• Cross-Functional Collaboration:
  
  explain forecasts, their assumptions and limitations to non-technical…