Nicolas Behar
Verified Expert in Engineering
Data Scientist and Developer
São Paulo - State of São Paulo, Brazil
Toptal member since February 3, 2026
Nicolas is a senior data scientist with over seven years of experience delivering high-impact machine learning solutions. He specializes in high-stakes areas such as credit risk, trading, and AML. Nicolas combines deep technical expertise in Python and Spark with a proven ability to scale production-ready architectures.
Portfolio
Experience
- Python - 8 years
- Machine Learning - 7 years
- Linear Regression - 7 years
- SQL - 6 years
- Natural Language Processing (NLP) - 5 years
- Forecasting - 4 years
- Credit Risk - 3 years
- Financial Engineering - 3 years
Preferred Environment
Ubuntu, Windows, Python, SQL, Scala, Slack
The most amazing...
...work I've performed for Demeter involved migrating a monolithic codebase to a modular system without incidents and a processing time reduction of 70%.
Work Experience
Data Science Tech Lead
Hapvida
- Developed a Bayesian Network model to map relevant clinical features via a conditional relative risk table and to determine if a patient is a high-risk patient that was misclassified by the radiologist, flagging it for a second review.
- Created a framework to evaluate X-Ray models made by vendors, being able to measure machine learning and operational performance on out-of-sample data, identify bias and shortcuts that the models could exploit (tubes and wirings).
- Standardized ML project development and evaluation on my team, and reduced time spent by data scientists on engineering tasks by creating bridges to the engineering and platform teams.
Quantitative Analyst
Demeter Agro Solutions
- Led and executed a critical infrastructure initiative to improve model training and backtest code, reducing by up to 70% the total processing time and adding more flexibility to test new strategies, models, and monitor features and models.
- Instilled a culture of profiling, Git versioning, and more detailed documentation to ensure code maintainability and process understanding.
- Elevated model risk discussions, focusing on model uncertainty, features, predictions, and residuals monitoring.
Decision Scientist
BS2 Bank
- Designed anti-money laundering (AML) workflows and alerting tools for foreign exchange internal and external customers and card payments and betting payment providers.
- Built an internal Python library to standardize AML alert creation, sharply increasing alert quality and enabling same-day deployment of multiple AML alerts. Reduced lead time from multiple days to hours.
- Partnered with AML stakeholders in the ideation, creation, and deployment of models and rules, ensuring all alerts met strict regulatory and internal risk requirements.
Data Scientist
Nubank
- Simplified customer lifetime value estimation for Nuconta (savings account), by developing a faster, more scalable approach that could be customized by other analysts with their views on future outcomes.
- Built and deployed a causal inference model to optimize marketing campaigns for a key product, targeting only persuadable users through nudges (SMS, email, push) and reducing attrition by avoiding customers not interested in the product.
- Supported Nuconta's AWS cost optimization companywide initiative by building and reviewing components of a forecasting model for critical services, effectively reducing in more than 50% the AWS costs for these services.
- Contributed to internal tooling, adding new functionalities for machine learning preprocessing tasks and causal meta-learners.
Data Scientist
Nu Invest Corretora de Valores
- Developed Spark-based data pipelines and infrastructure that streamlined machine learning workflows for the data science team.
- Leveraged NLP to extract actionable insights from chatbot interactions, driving cross-functional improvements that increased customer satisfaction.
- Forecasted support ticket volume using time-series modeling to avoid workforce shortage and reduce ticket resolution time.
Data Scientist
Guiabolso
- Contributed to the development and maintenance of credit scoring models, helping expand financial access for underserved Brazilians.
- Designed and implemented an automated collection framework from scratch, saving the team eight hours of manual labor every week.
- Designed a "share of wallet" analytical framework to benchmark Guiabolso's performance against key competitors in the retail credit market.
Experience
Volatility Forecasting
I held end-to-end responsibility for the project, including metric selection, LLM prompt engineering, and robust validation of the linear regression coefficients and residuals. To ensure reliability, I implemented conformal prediction to quantify model uncertainty.
Education
Master's Degree in Economics
Getulio Vargas Foundation - São Paulo, Brazil
Bachelor's Degree in Aerospace Engineering
Federal University of ABC (UFABC) - São Bernardo do Campo, Brazil
Bachelor's Degree in Science and Technology
Federal University of ABC (UFABC) - Santo André, Brazil
Skills
Libraries/APIs
Pandas, Scikit-learn, Matplotlib, GraphX
Tools
Slack, MATLAB
Languages
Python, SQL, Scala, Fortran, JavaScript
Storage
Data Pipelines
Paradigms
Unit Testing, Object-oriented Programming (OOP)
Platforms
Amazon Web Services (AWS), Ubuntu, Windows, Google Cloud Platform (GCP), Databricks
Frameworks
Spark
Other
Machine Learning, Model Monitoring, Feature Analysis, Linear Regression, Data Science, Finance, Data Analysis, Research, Data Analytics, Multivariate Statistical Modeling, Statistical Modeling, Financial Engineering, System Design, Programming, Mathematics, Physics, Forecasting, Natural Language Processing (NLP), Credit Risk, Data Wrangling, Software Engineering, Econometrics, Time Series, Time Series Forecasting, Equity, System Architecture, Derivative Pricing, Fixed-income Derivatives, Stochastic Differential Equations, Volatility, Wealth Management, Reliability Engineering, Aerodynamics, Algorithmic Trading, Investments & Risk Management, Data Warehousing, Risk Scorecards, A/B Testing, Large Language Models (LLMs), Stock Market, MLflow, Artificial Intelligence (AI), AI Project Management
How to Work with Toptal
Toptal matches you directly with global industry experts from our network in hours—not weeks or months.
Share your needs
Choose your talent
Start your risk-free talent trial
Top talent is in high demand.
Start hiring