Senior Data Scientist
Mutt Data
November 2025 - Present. Client: a leading Latin American beverage bottler.
- Set the build priorities of a causal evaluation engine for the client's commercial initiatives, ranking features by business value and technical complexity. The engine measures sales impact every two weeks for 1.5M+ customers (700k in Mexico and 800k in Brazil), with Matrix Completion for Causal Panels as the main model and Causal Forest as the benchmark.
- Estimate the impact of cross-company sales initiatives with causal forests, LightGBM-based causal models, synthetic control, difference-in-differences, and instrumental variables. Results were presented to the lead Growth Manager and escalated to the CEO.
- Design and analyze experiments to boost sales, customer digitalization, and satisfaction: treatment and control groups, sample sizes, sampling and randomization, and satisfaction surveys.
- Design demand elasticity models that combine microeconomics, causal inference, and large transactional databases.
- Build data pipelines and models in Databricks with PySpark, Spark SQL, and advanced SQL, using Databricks Connect, shared Python libraries, MLOps practices, Git, and Azure DevOps.
- Designed an agentic development workflow on Claude Code: an orchestrator delegates work to parallel subagents (one per Git worktree or Jira ticket) under human approval gates, supported by custom skills.
- Mentor two junior data scientists, interview candidates, and prioritize a high volume of stakeholder requests.
