Senior Data Scientist

Description

Summary: Seeking a Senior Data Scientist to lead the evolution of forecasting and marketing analytics capabilities, focusing on advanced time series models and business impact. Highlights: 1. Ownership of key forecasting and marketing analytics capabilities 2. Significant autonomy and technical ownership in modeling decisions 3. Translate complex models into actionable business insights and decisions **About the Role** ------------------ We are looking for a **Senior Data Scientist** to take ownership of a key part of our forecasting and marketing analytics capabilities. The main challenge will be to evolve our current regression\-based forecasting solution toward more advanced approaches, particularly **Transformer\-based architectures for time series forecasting**. This is not a role focused solely on implementing predefined models. We are looking for someone who can understand the business and data problem, define the modeling approach, experiment with alternatives, rigorously validate results, and ultimately help bring models into production. You will have significant autonomy and technical ownership, working closely with the team while bringing your own judgment to modeling and architectural decisions. **What You'll Do** ------------------ * Design, develop, train, and validate advanced **time series forecasting models**. * Evaluate **Transformer\-based forecasting architectures**, including **Chronos\-2**, as well as alternatives such as **TFT, PatchTST, Informer**, or other relevant approaches. * Build and run experiments using **Python, PyTorch, NumPy, Pandas, and Scikit\-learn**. * Train and validate models using **AWS SageMaker**. * Design rigorous validation strategies for time series, including: + Temporal validation + Backtesting + Rolling and expanding training windows + Leakage prevention + Evaluation on future/unseen periods * Compare different models and architectures, going beyond metrics to understand and explain why one approach performs better than another. * Translate forecasting results into **actionable business insights and marketing decisions**. * Contribute to **Marketing Mix Modeling (MMM), attribution, and budget optimization** initiatives. * Analyze channel contribution, incremental impact, saturation effects, and different investment scenarios. * Apply statistical inference, optimization, and model interpretability techniques to turn predictive outputs into business recommendations. * Clearly communicate technical decisions, assumptions, results, and trade\-offs to both technical and business stakeholders. **What We're Looking For** -------------------------- * Strong experience in **Machine Learning and Data Science**, with a particular focus on **time series and forecasting**. * Hands\-on experience building and validating forecasting models in real\-world scenarios. * Strong understanding of **time\-series validation methodologies** and the challenges associated with evaluating models on future data. * Strong proficiency in **Python**. * Experience with **PyTorch** and the scientific Python ecosystem, including **NumPy, Pandas, and Scikit\-learn**. * Experience working with **AWS SageMaker** for model training and experimentation. * Solid background in **statistics, statistical inference, optimization, and model interpretability**. * Ability to independently approach open\-ended problems, from data exploration and methodology selection to model validation and actionable conclusions. * Strong analytical judgment and the ability to justify modeling and architectural decisions. * Ability to connect technical results with business outcomes. **Nice to Have** ---------------- * Experience with **Transformer architectures for time series forecasting**, particularly **Chronos\-2**. * Experience with other forecasting architectures such as **Temporal Fusion Transformer (TFT), PatchTST, or Informer**. * Knowledge of **Bayesian Optimization** or tools such as **scikit\-optimize**. * Experience with **Marketing Mix Modeling (MMM), marketing attribution, or causal modeling**. * Understanding of concepts such as: + Adstock + Saturation curves + Channel effects + Incremental contribution + Budget allocation and optimization * Experience with tools or frameworks such as **Robyn, PyMC\-Marketing, or LightweightMMM**. **What Success Looks Like** --------------------------- You will be able to take a relatively open\-ended data problem and drive it from initial exploration to a validated solution. Success in this role is not just about achieving better model performance. It is about **choosing the right methodology, validating it correctly, explaining why it works, and translating the results into decisions that create measurable business value**. We're looking for someone who combines **strong technical expertise, rigorous experimentation, and business thinking** — and who is comfortable taking ownership of complex modeling challenges.