|
Описание: |
Role Type: Contract / Fixed-Scope Consulting Engagement
Duration: 2–3 months (10–12 weeks)
Allocation: Full-time or near full-time (4–5 days/week preferred)
Location: Remote (Must overlap with Spain / CET ± 6 hours)
Languages Required: Spanish (C1+) and English (B2+)
The Opportunity
We are looking for a hands-on Senior Data Science / Analytics Engineering Consultant to audit, redesign, and rebuild our core chargeback estimation and revenue estimation models.
This is a critical, end-to-end project. You will turn existing prototypes into production-grade pipelines within our modern data stack (dbt on Databricks & Python). Additionally, you will design and build a lightweight, business-facing scenario tool that enables finance and operations stakeholders to run “what-if” revenue simulations.
Our Technical Philosophy: Parsimony Over Black Boxes
We explicitly favor simple, highly interpretable, parametric approaches. If your first instinct for a forecasting problem is to throw an XGBoost ensemble, a deep learning model, or an LLM at it, this is not the project for you.
We build models where every parameter carries an explicit business meaning that can be explained to non-technical stakeholders in plain language (e.g., “If price increases by 10%, conversion drops by X% because our calculated price elasticity is −1.3”).
Scope of Work & Deliverables
* Phase 1: Discovery & Audit (Weeks 1–2): Review current estimation models (assumptions, dbt lineage, validation methods). Reconcile historical model outputs against actual accounting figures for the last 12–24 months to quantify bias and error. Deliver a written findings report.
* Phase 2: Redesign (Weeks 3–4): Propose the target parametric methodology for each model (granularity, refresh cadence, uncertainty quantification) and secure stakeholder sign-off.
* Phase 3: Rebuild (Weeks 5–9): Productionize the new models in dbt, Databricks, and Python. Use curve-based formulations with fitted historical parameters and clear confidence intervals. Implement strict unit/dbt testing and MLflow tracking.
* Phase 4: Scenario Tool (Weeks 9–10): Build a lightweight business interface (e.g., Streamlit on Databricks or a parameterized Databricks SQL dashboard). Users must be able to input hypotheses (price ±X%, volume ±Y%, mix shifts) and instantly view projected net impact with uncertainty bands.
* Phase 5: Handover (Weeks 11–12): Conduct knowledge transfer sessions, deliver runbooks/model cards, and provide a 2-week post-handover bug-fix support window.
Your Profile
* Experience: 5–8+ years in Data Science, Quant Analytics, or Senior Analytics Engineering.
* Production Ownership: Proven track record of owning financial-estimation or forecasting models in production end-to-end.
* The “Parsimony” Mindset: Demonstrable expertise in parametric/curve-based modeling (price elasticity, cohort survival curves, demand curves, GLMs) over black-box ML.
* Stakeholder Tooling: Experience building clean “what-if” tools or financial planning interfaces for business leadership.
* The Stack: Advanced Python (pandas, numpy, scikit-learn, statsmodels, and/or scipy.optimize), production dbt (incremental models, snapshots, tests), and Databricks (PySpark, Delta, workflows).
* Domain Knowledge: Direct experience in payment risk, chargebacks, refund/dispute modeling, revenue forecasting, cohort LTV, or pricing analytics.
* Communication: Exceptional ability to cross-reference data science outputs with strict financial accounting figures.
* Languages: Spanish (C1+) is mandatory for daily synchronization with local finance/ops teams; English (B2+) is required for technical documentation.
Nice to Have
* Background in FinTech, subscription models, payments, or high-volume e-commerce.
* Deep familiarity with Unity Catalog, Databricks Workflows, or Airflow.
* A strong track record of successful independent consulting engagements (references requested).
Відгукнутись на вакансію |
