Senior Data Scientist / Time Series Specialist (PoC & Forecasting)
About the Role
We are looking for a Senior Data Scientist to lead our advanced time series forecasting and accuracy-improvement initiatives. In this role, you will not just build models—you will drive end-to-end problem solving, from data-driven hypothesis formulation and hierarchical forecasting to local PoC verification.
If you thrive on untangling complex data structures, tackling seasonal patterns, and translating ambiguous business requirements into high-impact technical solutions, we’d love to have you on board.
Key Responsibilities & Scope
Primarily support in the following areas:
- Hypothesis Formulation: Drive accuracy improvements through rigorous, data-driven analysis. (Note: Grounded in data-backed investigation).
- Evaluation Frameworks: Design structured evaluation frameworks to determine improvement feasibility, delivering clear outputs that support PoC result interpretation and expected accuracy gains. (Includes scoping out the rationale behind PoC results).
- Error Isolation & Root Cause Analysis: Isolate accuracy-deteriorating segments and perform deep-dive cause analysis to answer "why accuracy is dropping."
- Impact-Driven Prioritization: Prioritize improvement candidates based on expected business impact and implementation feasibility, selecting optimal items for PoC verification.
- Local Environment Verification (PoC Level): Conduct robust testing and verification in a local environment. (Note: Production deployment is handled separately by the customer team).
What You Bring (Technical Expertise)
1. Time Series Forecasting Mastery
- Classical & ML Approaches: Deep knowledge of classical methods (ARIMA/SARIMA, ETS, Prophet) and modern ML approaches (XGBoost/LightGBM with lag features).
- Hierarchical Forecasting: Experience with multi-level forecasting (SKU → style → category → region/channel) and reconciliation methods.
- Complex Patterns: Proven ability to handle complex seasonality (weekly/monthly), trends, holidays, and promotional effects.
2. Data Quality & Advanced Testing
- Feature Engineering: Crafting robust features (lags, rolling statistics, price/promo features, calendar/holiday signals, stock-outs, lead times).
- Evaluation & Backtesting: Selecting appropriate metrics (WAPE, MAPE, sMAPE, RMSE) and designing backtesting strategies (rolling/expanding window).
- Model Insights: Strong grasp of bias/variance tradeoffs, regularization, and model interpretability (using tools like SHAP).
3. Tech Stack & Data Wrangling
- Advanced SQL & handling large-scale datasets.
- Python Ecosystem: Expert-level proficiency in pandas, numpy, scikit-learn, statsmodels, prophet, and tree-based models (XGBoost, etc.).
- Cloud & Tools: Experience with cloud environments (GCP), workflow orchestration (Airflow), and data warehousing (BigQuery).
Core Competencies & Engineering Culture
We value autonomy, high ownership, and proactive collaboration:
- Strategic Communication & Ownership: You don’t wait passively for clean requirements. You lead discussions, ask sharp clarifying questions, align stakeholders on priorities, and communicate risks early.
- Deadline & Delivery Reliability: You take ownership of timelines, set realistic scopes, and ensure complete transparency through regular updates and proactive early escalation if risks arise.
- Systematic Problem Solving & Catch-up: Ability to rapidly digest complex, undocumented business systems, interview stakeholders to uncover hidden issues, and turn fragmented information into structured action items.
- Quality Assurance & Craftsmanship: Active code-review participant, independent test-case designer (happy paths, edge cases, boundaries), and an advocate for clean documentation (design docs, runbooks, incident logs).
Why Join Us?
- Work on challenging, high-impact forecasting problems with cutting-edge data stacks.
- High degree of autonomy and direct impact on business accuracy and performance.
- A collaborative culture that values technical rigor, clean code, and intellectual curiosity.