About Babs
Economics taught me to ask why. Data science gave me tools to test what works.
I build analytical systems for decisions that affect people, institutions, and markets—from health financing in Rwanda to pricing systems in Europe and digital-health tools across emerging markets.
The through-line
01I care about the distance between a technically correct answer and a genuinely useful one.
The best model is rarely the most complicated model. It is the one built around the decision, honest about uncertainty, and clear enough for people to challenge and use.
My path began in mathematics and quantitative economics, moved through political-violence research and business intelligence, and developed into applied machine learning across healthcare, pricing, recommendations, and experimentation.
That breadth shapes how I work today: start with the institutional context, build the minimum credible evidence, and design the output around action.
Career path
02Data Scientist, Causal Foundry
Building claims-driven systems for health financing, provider performance, prescribing quality, anomaly detection, and operational monitoring.
Senior Data Scientist, ZF Group
Developed pricing-optimization, elasticity, uplift, and anomaly-detection systems for commercial decisions across a large industrial portfolio.
Data Scientist, Benshi AI
Built demand forecasts, survival models, recommender systems, and adaptive experiments for healthcare workers and pharmacy networks.
Business intelligence and political-violence research
Worked on customer analytics and decision dashboards after building newspaper-text workflows for research on conflict risk at IAE-CSIC.
MSc Quantitative Economics
Studied at Université Paris 1 Panthéon-Sorbonne and Universitat Autònoma de Barcelona as an Erasmus Mundus scholar.
How I work
03Start with the decision
I define who acts, what changes, and what a useful answer must contain before reaching for a model.
Make uncertainty visible
I separate signal from confidence, document assumptions, and build checks that invite scrutiny rather than hide it.
Design for adoption
I translate across research, engineering, policy, and product teams so evidence can move into real workflows.
Toolkit
04Research & decisions
Causal inference · A/B testing · survival analysis · forecasting · Bayesian modeling · decision theory
Machine learning
NLP · recommender systems · anomaly detection · pricing optimization · multimodal learning · reinforcement learning
Data & production
Python · SQL · PySpark · BigQuery · Databricks · MLflow · Airflow · Google Cloud · AWS
The short version