Alice

- Predictive Health Risk Model: Led the development of a predictive model that assesses new customers' risk of intensive health system usage. Collaborated closely with Operations to implement proactive interventions, reducing uncoordinated usage among high-risk individuals. This solution achieved a 15% cost reduction in high-risk segments while ensuring optimal health outcomes.
- OpenAI-Powered Chat Solution: Contributed to the creation of a two-agent chatbot system that assists customers with health-related inquiries. One agent gathers and summarizes the user's needs, while the second analyzes protocols and recommends appropriate healthcare professionals. This innovation, developed in partnership with the Health Product team, reduced the time to resolution by 20% without compromising customer experience or health outcomes.
- Data Science Python Library Refactor: Spearheaded a comprehensive refactor of the Data Science Python library used in production environments (Airflow DAGs, FastAPI apps). Optimized the codebase for integration with external services (S3, Redshift, OpenAI, Google Sheets), enhanced test coverage to near 100%, and added asynchronous capabilities, resulting in a significant reduction in development time for the Data Science team.

Senior Data Scientist

Inventa

- Build, maintain, and improve the ML credit limit service, and the real-time fraud detection system, with hundreds of daily requests, avoiding multiple malicious transactions and managing Inventa Marketplace's risk exposure.
- Design and implement Python (scikit-learn) based ML applications (hosted on Wyvern AI and Modelbit), using either REST or Webhook API designs; powered by an internal Feature Store solution (based on Feast, Snowflake, and Redis); and whose observability is handled by Dbt (data quality), Graphana (Monitoring) and Evidently AI (Observability)
- Supervise the implementation of the Experimentation Infrastructure (Eppo) and share best practices for experimentation within the Data team, along with executing the ones within my domain (Risk) which generated 17% of extra GMV for retained users (around 40% of the user base).

Senior Data Scientist - Risk

Wildlife Studios

- Implement probabilistic techniques into a statistical PySpark library, including bootstrapping, statistical tests, variance reduction techniques, Bayesian techniques, and Peak Proof inference, responsible for powering the company-wide experimentation platform, which executed dozens of simultaneous experiments, with hundreds of thousands of users each, leading, in conjunction, to an estimated 23% of extra ARPU on new games (around 15% of the user base).
- Be the company-wide causal inference sponsor, standardizing experimentation & and causal inference best practices, and promoting the experimentation culture across the organization

Data Scientist - Experimentation Platform

- Work within a core game (Zooba) Product team to develop analyses and models to assess or create new hypotheses of game features or changes, and measure their efficiency through different experiment designs. Responsible for deploying at least 3 new features, and avoiding other 4 harmful or neutral changes.
- Build models, dashboards, reports, and data pipelines based on Airflow, and model and analyses powered by the Databrics ecosystem (PySpark and MLflow)

Data Scientist - Product Insights

Uber

- Analyze and track real-time pricing algorithms, driver movement, and driver preferences products, along with managing its experimentation and deployment, across the LatAm territory; overseeing the test and deployment of 3 pricing features
- Recommend driver incentives allocation decisions through statistical analysis for key regional operations, such as the incentives strategy in the fierce competitive landscape of 2022 End-of-year driver supply in Sao Paulo.
- Design and analyze experiments, using classical AB tests or special frameworks such as Switchback or Synthetic Controls, using Uber’s internal tech stack, mostly based on Python, Hive, Presto, and Spark;

LatAm Central Marketplace Ops - Driver Pricing Manager

Work within the LatAm Driver Experience team to develop analyses and models to segment drivers' behavior, and comprehend their engagement, and satisfaction with the platform. Responsible for assessing the neutral effect of a costly regional effort to engage drivers physically, and implement a new policy to include estimated hourly net revenue from drivers into Driver Pricing policies.
- Models, analysis, and dashboards powered by Uber’s internal tech stack, mostly based on Python, Hive, Presto and Spark; along with experimentation through their internal Exp Platform (Mentana)

LatAm Central Driver Ops - Driver Analytics 

- Responsible for analytics and data related stuff within Central Driver Experience Team.
- Main Central Driver Operations dashboards owner, analytics reports executor and personal development assistance for team members focused on analytics and coding skills.

LatAm Central Driver Ops  - Driver Experience Intern

Força Aérea Brasileira - FAB

- Military Service, focused on the Brazilian Air Force Officers formation, aiming to develop leadership, discipline, respect, organisation and results.
- Training on basic military competencies, decision making at high pressure situations, and other characteristics requested to be a future air force pilot.

Cadet at "Escola Preparatória de Cadetes do Ar"

Data Science, Experimentation & Causal Inference

- Data Scientists with 6 years of experience in Analytics, Experimentation and Data Modeling
- 4 years of experience deploying Machine Learning models into Production on different industries
- 1 year of experience deploying LLM based applications in Health triage
- Causal Inference specialist with 2 large real life projects delivered (Experimentation Platform and Credit Policy Allocator)

Taplio

Rafael Bicudo Rosa's Linkedin Analytics

Rafael Bicudo Rosa

Want to drive more opportunities from LinkedIn?