Senior Data Science and AI Leader

Took a career break for family reasons.

Led the development of machine-learning pricing models, including Tweedie GLM models, underwriting systems, AI-driven decision engines, and risk analytics for a top Brazilian health insurance carrier. Translated business requirements into scalable AI and pricing solutions and presented model outcomes to the board.

Managed three FTEs and coordinated five consultants while building data partnerships with Serasa Experian. Led competitor and KPI analysis that informed regional product redesign.

Deployed deep-learning models on GCP using Python, reducing the loss ratio by 1.4 percentage points, equivalent to approximately €6 million annually. Led stakeholder management across Risk, Pricing, and AI teams.

Led cross-functional teams of up to 15 professionals across Data Science, Data Engineering, and MLOps at an AI startup. Coordinated multiple analytics and AI product initiatives while reporting to the Executive Director.

Led the end-to-end launch of an AI customer-segmentation product for leading Brazilian telecommunications companies. Used machine learning, SQL, and Power BI to support marketing campaigns that increased campaign ROI by 15%.

Designed segmentation software using XGBoost, Logistic Regression, and Random Forest in Python on AWS. Implemented GenAI call-auditing solutions using LLMs, RAG, ChatGPT, vector databases, and prompt engineering, reducing manual QA effort by 30%.

Directed financial-crime analytics and machine-learning initiatives for trading surveillance, using Python, SAS, and SQL. Delivered regulatory KPIs and Brazilian Securities Commission compliance reporting through Tableau.

Built and led a new analytics team of four data and decision scientists, hiring and mentoring the team from scratch. Served as the risk-product gatekeeper for new financial instruments, defining controls, thresholds, trading limits, and exposure parameters.

Deployed an AI anomaly-detection surveillance model covering 50 million weekly transactions, reducing operational workload by 80%. Built executive risk dashboards for real-time product controls and security-threshold governance.

Relocated to Dublin, Ireland during the COVID-19 pandemic. Resumed professional work remotely with B3 in April 2021.

Led end-to-end fraud detection and anti-money-laundering initiatives for a Santander Group acquiring bank. Developed real-time transaction-monitoring capabilities using SAS, Python, machine learning, and feature engineering.

Built a cross-functional team of five data scientists and fraud analysts. Defined fraud policy with Finance, Risk, and executive leadership, ensured Central Bank compliance, and advised key e-commerce clients on fraud-product implementation and rule tuning.

Deployed real-time ML fraud models and alerting rules that supported 40% transaction growth while reducing fraud losses by 10% year over year.

Developed PD credit-scoring and fraud-scoring models using SQL, SAS, and Logistic Regression for corporate, SME, and retail portfolios. Optimized collection policies, fraud alerts, detection models, and business rules across Risk, Fraud, and Operations.

Acted as a technical lead for financial-product evaluation, profitability analysis, CRM analytics, and integrated risk management across issuing and acquiring operations.

Delivered credit and fraud models with KS of 47%; the credit score increased market penetration by 25% with minimal losses. Reduced fraud losses by 30%, identified a multi-million-euro internal fraud scheme, and built a Balanced Scorecard that increased case-analysis productivity by 30%.

Developed PD credit and behavioral scoring models for B2B and B2C telecom portfolios using SAS and SPSS. Calculated credit limits, optimized credit policies, and forecast financial outcomes.

Led credit analysts and supported credit-risk decision making for a US telecommunications operator.

Reduced the rejection rate from 20% to 10% through the first credit-score models without affecting involuntary churn.

Developed customer clustering, segmentation, forecasting, targeting, and campaign analytics for an Italian e-commerce wine retailer serving more than six million customers in the EU market.

Applied k-means segmentation and A/B testing to improve CRM and marketing performance.

Increased sales by 15% through customer segmentation initiatives.

Master's degree in Analysis and Evaluation of Complex Systems, with a focus on Economic Statistics.

European Union Erasmus Programme.

Bachelor's degree in Statistics.