Lucas is available for hire
Hire LucasLucas Cavalcante
Verified Expert in Engineering
Data Scientist and Developer
Aracaju - State of Sergipe, Brazil
Toptal member since June 3, 2026
Bio
Lucas is a senior data scientist and consultant specializing in AI, machine learning, and operations research. With a master's degree in applied mathematics, he transforms complex data into strategic solutions, helping organizations make better decisions through advanced analytics and data-driven methodologies.
Portfolio
Natura Brazil
Python, Pandas, SciPy, Databricks, PySpark, StatsModels, Machine Learning...
3Blue Media
Python, Pandas, MongoDB, SQL, Slack, LightGBM, StatsModels, NumPy...
Vale
Analytics, Apache Airflow, Kubeflow, Docker, TensorFlow, BERT...
Experience
- Data Analytics - 6 years
- SciPy - 6 years
- Python - 6 years
- StatsModels - 6 years
- Artificial Intelligence (AI) - 6 years
- Data Science - 6 years
- SQL - 6 years
- Machine Learning (ML) APIs - 6 years
Preferred Environment
Data Science, Data Analytics
The most amazing...
...project I've built combined operations research and ML to optimize SKUs and cold-start allocation for 400+ stores, boosting revenue by 4%.
Work Experience
Senior Data Science
2026 - PRESENT
Natura Brazil
- Led the eCommerce demand forecasting strategy at a national level, developing 30-day predictive models that optimized factory manufacturing schedules and significantly reduced stockouts and inventory surpluses. The error reduction is 357 pp.
- Managed time-series and seasonality analysis to accurately project future order volumes, guiding the strategic recruitment of temporary and extra factory staff during peak demand periods.
- Structured predictive models based on historical sales data and top 10 product performance, streamlining operational capacity and driving a positive impact on the company's overall revenue.
Technologies: Python, Pandas, SciPy, Databricks, PySpark, StatsModels, Machine Learning, NumPy, Regex, SQL, Jupyter Notebook, Matplotlib, Seaborn, Jira, Data Science, Data Analytics, Data Engineering, Presentations, Scikit-learn, Feature Engineering, Regression Modeling, Linear Regression, Pricing Elasticity, Statistics, Azure, Time Series Forecasting, Git, Data Analysis, Azure Databricks, Bayesian Statistics, Data Pipelines, Demand Forecasting, Model Deployment, Model Development, Model Evaluation, Monitoring, Random Forests, BigQuery, MLflow, Delta Lake, Time Series, Data Scientist, Large Language Models (LLMs), JSON, Data Quality, Azure Machine Learning, Jobs, Price Analysis, Dashboards, Hypothesis Testing, Supply Chain, Sprints, Canvas, AI Compliance, Pricing Models, Google Sheets, Key Performance Indicators (KPIs), Metrics, XGBoost, LightGBM, ETL, Google Cloud Platform (GCP), Machine Learning (ML) APIs, Data Visualization, Predictive Modeling, Data Validation, Cross-functional Coordination, Feature Prioritization, Product Ownership, Feature Planning, Product Management, Data Reporting, User Feedback, Prompt Engineering, Claude, Claude API, Communication, Stakeholder Management, ChatGPT, Low Latency, AI Evaluation, Statistical Analysis, Forecasting
Senior Data Scientist
2025 - 2026
3Blue Media
- Developed and implemented pricing elasticity frameworks using machine learning and advanced heuristic models to optimize pricing strategies and enhance profit margins.
- Architected the automation and migration of legacy pricing structures into scalable data pipelines, ensuring higher processing efficiency and robust model deployment.
- Engineered data-driven heuristic solutions to dynamically assess market behaviors, directly supporting strategic corporate decision-making and revenue optimization.
Technologies: Python, Pandas, MongoDB, SQL, Slack, LightGBM, StatsModels, NumPy, Jupyter Notebook, Visual Studio Code (VS Code), Causal Inference, A/B Testing, Backtesting, XGBoost, NeuralProphet, Metrics, Key Performance Indicators (KPIs), Regex, Google Sheets, Pricing Models, Pyomo, Feature Engineering, Regression Modeling, Linear Regression, Pricing Elasticity, Statistics, Machine Learning, Azure, Time Series Forecasting, Git, Data Engineering, Data Analysis, Bayesian Statistics, Data Pipelines, Deep Learning, Demand Forecasting, Model Deployment, Model Development, Model Evaluation, Monitoring, Random Forests, BigQuery, Time Series, Data Scientist, Large Language Models (LLMs), Fraud Detection, Fraud Prevention, JSON, Data Quality, Azure Machine Learning, Docker, Jobs, Price Analysis, Dashboards, Web Marketing, Hypothesis Testing, Kanban, NoSQL, AI Compliance, ETL, Natural Language Processing (NLP), Machine Learning (ML) APIs, Microsoft Power BI, Data Visualization, Predictive Modeling, Data Validation, Cross-functional Coordination, Feature Prioritization, Product Ownership, Feature Planning, Data Reporting, Prompt Engineering, Claude, Claude API, Communication, Stakeholder Management, APIs, ChatGPT, Low Latency, AI Evaluation, Pricing Optimization, Statistical Analysis, Forecasting
Senior Data Scientist
2024 - 2025
Vale
- Developed and deployed a BERT-based multi-class text classification model for compliance auditing, automating the categorization of documents into complaints, reports, and over 30 distinct legal and behavioral typologies.
- Automated the regulatory document review workflow in the compliance sector, eliminating manual processing and human classification errors while increasing operational efficiency by over 80%.
- Standardized legal inquiry classification data pipelines, significantly accelerating processing speeds and delivering a faster, more reliable compliance monitoring system for corporate auditing.
Technologies: Analytics, Apache Airflow, Kubeflow, Docker, TensorFlow, BERT, Natural Language Processing (NLP), Google APIs, Google Cloud Machine Learning, Vertex AI, Cryptography, PostgreSQL, MongoDB, RabbitMQ, Pandas, NumPy, SpaCy, Natural Language Toolkit (NLTK), Regex, Data Engineering, Data Analytics, Data Science, StatsModels, SciPy, Matplotlib, Seaborn, Scikit-learn, Tf-idf, Word2Vec, Metrics, AI Compliance, OpenAI, AutoML, Feature Engineering, Regression Modeling, Linear Regression, Statistics, Machine Learning, Azure, Databricks, Time Series Forecasting, PySpark, Git, Data Analysis, Data Pipelines, Deep Learning, Model Deployment, Model Development, Model Evaluation, Monitoring, Random Forests, BigQuery, MLflow, Google Cloud Storage, Delta Lake, Time Series, Data Scientist, Large Language Models (LLMs), Fraud Prevention, JSON, Amazon Web Services (AWS), Data Quality, Azure Machine Learning, Jobs, Dashboards, Hypothesis Testing, Key Performance Indicators (KPIs), XGBoost, LightGBM, Slack, ETL, Jira, Google Cloud Platform (GCP), Machine Learning (ML) APIs, Data Visualization, Predictive Modeling, Data Validation, Cross-functional Coordination, Feature Prioritization, Product Ownership, Feature Planning, Product Management, Data Reporting, User Feedback, PyTorch, Hugging Face Transformers, Sequence Models, Transformers, Google BigQuery, Sentiment Analysis, Speech-to-Text (STT), Prompt Engineering, Communication, Generative Artificial Intelligence (GenAI), APIs, ChatGPT, Vertex, Reinforcement Learning, Low Latency, AI Evaluation, Statistical Analysis
Senior Data Scientist
2024 - 2025
Lojas Riachuelo
- Developed and deployed an operational research model to optimize SKU distribution from centers to 400+ stores, leveraging probability of sales and logistics constraints to maximize fill rate and achieve a 4% revenue increase in A/B testing.
- Designed and implemented a recommendation system for allocating new SKUs with no sales history, utilizing clustering techniques to predict initial stock distribution, reducing inventory breakage and surplus by 50%.
- Integrated advanced statistical simulations for stockout forecasting within the supply chain, enabling dynamic replenishment planning, mitigating distribution disruptions, and preventing lost sales opportunities.
Technologies: Python, CatBoost, Pyomo, XGBoost, A/B Testing, Backtesting, NumPy, Pandas, Jupyter Notebook, Databricks, Azure Data Factory (ADF), StatsModels, Causal Inference, SciPy, Matplotlib, Seaborn, LightGBM, Heuristics, Statistical Methods, Modeling, Price Analysis, Jobs, Dashboards, Feature Engineering, Regression Modeling, Linear Regression, Pricing Elasticity, Statistics, Machine Learning, Azure, Time Series Forecasting, PySpark, Git, Data Engineering, Plotly, Data Analysis, Azure Databricks, Bayesian Statistics, Data Pipelines, Demand Forecasting, Model Deployment, Model Development, Model Evaluation, Monitoring, Random Forests, BigQuery, MLflow, Delta Lake, Time Series, Data Scientist, Large Language Models (LLMs), JSON, Data Quality, Azure Machine Learning, Docker, Hypothesis Testing, Supply Chain, Kanban, Sprints, Canvas, AI Compliance, Google Sheets, Key Performance Indicators (KPIs), Metrics, Slack, ETL, Jira, Regex, Machine Learning (ML) APIs, Data Visualization, Predictive Modeling, Data Validation, Cross-functional Coordination, Feature Prioritization, Product Ownership, Feature Planning, Product Management, Data Reporting, User Feedback, Prompt Engineering, Communication, APIs, ChatGPT, Vertex, Low Latency, AI Evaluation, Statistical Analysis, Forecasting
Senior Data Scientist
2023 - 2024
Grupo Boticario
- Developed a mathematical framework to optimize SKU distribution across 3,700+ retail stores, balancing inventory constraints to improve fill rates and significantly reduce stockouts.
- Conducted rigorous A/B testing and hypothesis testing to evaluate inventory allocation strategies, identifying statistically significant improvements in retail product availability.
- Engineered data-driven pipeline migrations and automation workflows for SKU management, integrating advanced anomaly detection to ensure inventory data quality and efficiency.
Technologies: Python, Pandas, Google Cloud Platform (GCP), NumPy, A/B Testing, Backtesting, StatsModels, SciPy, Matplotlib, Seaborn, SQL, Scikit-learn, Causal Inference, Linear Programming, Pyomo, Feature Engineering, Regression Modeling, Linear Regression, Pricing Elasticity, Statistics, Machine Learning, Time Series Forecasting, Git, Plotly, Data Analysis, Data Pipelines, Demand Forecasting, Model Deployment, Model Development, Model Evaluation, Monitoring, Random Forests, BigQuery, Google Cloud Storage, Delta Lake, Time Series, Data Scientist, JSON, Data Quality, Jobs, Hypothesis Testing, Supply Chain, Kanban, Sprints, Canvas, AI Compliance, Google Sheets, Key Performance Indicators (KPIs), Metrics, Slack, ETL, Jira, Regex, Machine Learning (ML) APIs, Data Visualization, Predictive Modeling, Data Engineering, Data Validation, Cross-functional Coordination, Feature Prioritization, Data Reporting, Google BigQuery, Communication, Vertex, Statistical Analysis, Forecasting
Database Marketing Manager
2021 - 2022
Claro
- Developed and deployed predictive machine learning models to identify customer churn, enabling proactive retention strategies and personalized interventions to improve customer loyalty.
- Applied unsupervised learning techniques, including K-Means and hierarchical clustering, to segment customers based on behavioral data, optimizing targeted marketing campaigns and funnel channels.
- Designed and executed controlled A/B tests and hypothesis testing frameworks to rigorously measure the impact of retention and marketing campaigns, ensuring statistically valid business conclusions.
Technologies: Python, StatsModels, SciPy, Causal Inference, Backtesting, A/B Testing, Scikit-learn, Jupyter Notebook, Pandas, NumPy, Visual Studio Code (VS Code), Matplotlib, Seaborn, Churn Analysis, Web Marketing, Hypothesis Testing, Regression Modeling, Statistics, Machine Learning, Time Series Forecasting, Git, Data Analysis, Model Deployment, Model Development, Time Series, Data Scientist, Kanban, Tableau, AI Compliance, Google Sheets, Key Performance Indicators (KPIs), Metrics, ETL, Jira, Machine Learning (ML) APIs, Microsoft Power BI, Data Visualization, Predictive Modeling, Cross-functional Coordination, Data Reporting, User Feedback, Communication, Statistical Analysis
Data Science Trainee
2019 - 2021
Senai - PR
- Developed time-series forecasting models using LSTM and ARIMA architectures to predict chicken weight and silo feed flow, optimizing slaughter dates and logistics scheduling while reducing operational waste.
- Designed and deployed deep learning solutions, including CNN and YOLO models, for real-time traffic sign detection and keystroke biometrics dynamics to enhance automated systems and security layers.
- Engineered an advanced credit scoring system for an agricultural cooperative using predictive analytics, successfully reducing default risks and tailoring financial profiles for members.
- Built an intelligent medical OCR auditing system and a KPI/OKR management chatbot, automating data extraction and streamlining decision-making for executive monitoring.
Technologies: Python, Pandas, NumPy, Scikit-learn, TensorFlow, Keras, StatsModels, SciPy, Matplotlib, Microsoft Power BI, Dashboards, Long Short-term Memory (LSTM), Churn Analysis, Fraud Audits, Fraud Prevention, Optical Character Recognition (OCR), Graph Theory, Algorithmic Graph Theory, MongoDB, Chatbots, Credit Scores, Feature Engineering, Regression Modeling, Linear Regression, Pricing Elasticity, Statistics, Machine Learning, Azure, Time Series Forecasting, PySpark, Git, PostgreSQL, Data Engineering, Selenium, Plotly, Data Analysis, Data Pipelines, Deep Learning, Demand Forecasting, Model Deployment, Model Development, Model Evaluation, Monitoring, Random Forests, BigQuery, MLflow, LSTM, Google Cloud Storage, Delta Lake, Time Series, Data Scientist, Fraud Detection, JSON, Data Quality, Jobs, Price Analysis, SpaCy, Natural Language Toolkit (NLTK), Hypothesis Testing, Scrum, Kanban, SQLite, NoSQL, Sprints, Canvas, Tableau, Azure Databricks, AI Compliance, Word2Vec, Pricing Models, Google Sheets, Key Performance Indicators (KPIs), Metrics, XGBoost, LightGBM, ETL, Regex, Natural Language Processing (NLP), Machine Learning (ML) APIs, Data Visualization, Predictive Modeling, Data Validation, Cross-functional Coordination, Feature Prioritization, Product Ownership, Feature Planning, Product Management, Data Reporting, User Feedback, Google BigQuery, Communication, AI Evaluation, Statistical Analysis, Forecasting
Experience
Optimizing Retail SKU Distribution
Developed and deployed an end-to-end supply chain optimization engine on Databricks to transform SKU distribution across 400+ retail stores.
The system integrates a CatBoost machine learning model for high-accuracy time-series demand forecasting, capturing complex seasonal trends and pricing elasticities. To model product scarcity and predict stockout frequencies under volatile retail conditions, the engine uses a probabilistic framework based on a Poisson distribution. These dynamic demand distributions feed directly into a mathematical optimization model built using Pyomo. The operational research engine solves high-dimensional allocation problems under strict logistics constraints, prioritizing store shipments based on real-time sales probabilities.
Validated through rigorous historical backtesting and a comprehensive live A/B testing framework, the solution successfully mitigated inventory surpluses and minimized stockout disruptions. This automated data pipeline is projected to drive a minimum annual revenue increase of 3%, establishing a highly scalable, data-driven framework for retail inventory management.
The system integrates a CatBoost machine learning model for high-accuracy time-series demand forecasting, capturing complex seasonal trends and pricing elasticities. To model product scarcity and predict stockout frequencies under volatile retail conditions, the engine uses a probabilistic framework based on a Poisson distribution. These dynamic demand distributions feed directly into a mathematical optimization model built using Pyomo. The operational research engine solves high-dimensional allocation problems under strict logistics constraints, prioritizing store shipments based on real-time sales probabilities.
Validated through rigorous historical backtesting and a comprehensive live A/B testing framework, the solution successfully mitigated inventory surpluses and minimized stockout disruptions. This automated data pipeline is projected to drive a minimum annual revenue increase of 3%, establishing a highly scalable, data-driven framework for retail inventory management.
Credit Score
At a major credit cooperative, I led the development of an end-to-end predictive modeling system to drive growth while mitigating financial risk. By analyzing historical customer data, the machine learning models identify high-potential prospects and dynamically estimate optimal credit limits. Simultaneously, advanced classification algorithms predict default probabilities to prevent bad debt. This balanced framework safely scales the portfolio, optimizing risk-adjusted returns and streamlining internal underwriting.
LATAM eCommerce
As a lead data professional at Natura, I spearheaded predictive modeling of eCommerce order volumes across LATAM, directly impacting factory-floor operations and raw-material procurement. To handle highly diverse product lifecycles, I designed and deployed segmented LightGBM models. I used a standard framework for high-volume items with robust histories, implemented a Tweedie objective to handle intermittent, sparse historical data, and leveraged a Poisson distribution to naturally accommodate the positive skew typical of supply chain metrics.
By balancing statistical accuracy (WAPE) with critical business KPIs such as overstock and stockout rates, the solution ensured lean manufacturing without compromising product availability. Following rigorous A/B testing on randomized SKU groups to validate operational impact, I collaborated with DevOps to build automated, production-ready pipelines. Utilizing Azure Data Factory for job orchestration and scheduling, alongside MLflow for model tracking and governance, the system now runs continuously in production, successfully transforming raw internal data into actionable supply chain intelligence.
By balancing statistical accuracy (WAPE) with critical business KPIs such as overstock and stockout rates, the solution ensured lean manufacturing without compromising product availability. Following rigorous A/B testing on randomized SKU groups to validate operational impact, I collaborated with DevOps to build automated, production-ready pipelines. Utilizing Azure Data Factory for job orchestration and scheduling, alongside MLflow for model tracking and governance, the system now runs continuously in production, successfully transforming raw internal data into actionable supply chain intelligence.
Education
2019 - 2020
Professional Degree in Artificial Intelligence
National Industrial Training Service - Londrina, Paraná, Brazil
2016 - 2018
Master's Degree in Applied Mathematics and Computation
State University of Londrina - Londrina, Paraná, Brazil
2015 - 2016
Bachelor's Degree in Mathematics and Computer Science
Federal University of Sergipe - Aracaju, Sergipe, Brazil
2006 - 2012
Bachelor's Degree in Mathematics
Federal University of Sergipe - Aracaju, Sergipe, Brazil
Certifications
JULY 2026 - JULY 2028
Partner Training - Data & AI Governance with Unity Catalog
Databricks
JULY 2026 - PRESENT
Knowledge Badge - Machine Learning Model Deployment
Databricks
JULY 2026 - PRESENT
Knowledge Badge - Data Preparation for Machine Learning
Databricks
JULY 2026 - PRESENT
Knowledge Badge - Machine Learning Model Development
Databricks
Skills
Libraries/APIs
Scikit-learn, TensorFlow, SciPy, PySpark, Claude API, Pandas, Keras, Matplotlib, NumPy, SpaCy, Natural Language Toolkit (NLTK), LSTM, PyTorch, Hugging Face Transformers, OpenCV, XGBoost, CatBoost, Google APIs
Tools
StatsModels, Claude, Microsoft Power BI, Azure Machine Learning, Apache Airflow, Seaborn, AutoML, Git, Plotly, BigQuery, ChatGPT, MATLAB, Canvas, Tableau, Scilab, You Only Look Once (YOLO), Jira, Slack, Google Sheets, RabbitMQ
Languages
Python, SQL, Regex
Platforms
Databricks, Google Cloud Platform (GCP), Kubeflow, Docker, Azure, Jupyter Notebook, Visual Studio Code (VS Code), Vertex AI, Amazon Web Services (AWS)
Storage
Data Validation, NoSQL, SQLite, JSON, PostgreSQL, Data Pipelines, Google Cloud Storage, MongoDB
Frameworks
LightGBM, Selenium, Streamlit
Paradigms
ETL, Scrum, Kanban, Linear Programming
Industry Expertise
Insurance
Other
Data Science, Artificial Intelligence (AI), Data Analytics, Machine Learning, Machine Learning (ML) APIs, Data Engineering, Time Series Forecasting, Linear Regression, Regression Modeling, Data Analysis, Data Scientist, Data Visualization, Predictive Modeling, Cross-functional Coordination, Data Reporting, Prompt Engineering, Communication, Analytics, Optical Character Recognition (OCR), Fraud Detection, Credit Scores, BERT, Natural Language Processing (NLP), Pricing Models, Tf-idf, Word2Vec, AI Compliance, OpenAI, Web Marketing, Hypothesis Testing, Fraud Prevention, Supply Chain, Feature Engineering, Pricing Elasticity, Statistics, Azure Databricks, Bayesian Statistics, Deep Learning, Demand Forecasting, Model Deployment, Model Development, Model Evaluation, Monitoring, Random Forests, MLflow, Delta Lake, Time Series, Large Language Models (LLMs), Data Quality, Credit Risk, Feature Prioritization, Product Ownership, Feature Planning, Product Management, User Feedback, Sequence Models, Transformers, Google BigQuery, Generative Artificial Intelligence (GenAI), APIs, Vertex, Reinforcement Learning, Low Latency, AI Evaluation, Pricing Optimization, Statistical Analysis, Forecasting, Sprints, Applied Mathematics, Discrete Mathematics, Linear Algebra, Partial Differential Equations, Differential Equations, Linear Optimization, Generalized Linear Model (GLM), Optimization, Data Structures, Modeling, Programming, Multivariate Calculus, Quantitative Calculus, Calculus, Physics, Graph Theory, Algorithmic Graph Theory, Heuristics, Fraud Audits, Long Short-term Memory (LSTM), Convolutional Neural Networks (CNNs), Churn Analysis, Presentations, Causal Inference, A/B Testing, Backtesting, NeuralProphet, Metrics, Key Performance Indicators (KPIs), Pyomo, Azure Data Factory (ADF), Statistical Methods, Price Analysis, Jobs, Dashboards, Google Cloud Machine Learning, Cryptography, Chatbots, Multivariate Statistical Modeling, Probability Theory, eCommerce, Pricing, Hyperparameter Tuning, Model Tracking, Feature Management, Data Preparation, Data Cleaning, Batch Inference, Pipeline Deployment, Real-time Deployment, Performance Optimization, Governance, Proof of Concept (POC), Architecture, Sentiment Analysis, Speech-to-Text (STT), Stakeholder Management
Collaboration That Works
How to Work with Toptal
Toptal matches you directly with global industry experts from our network in hours—not weeks or months.
1
Share your needs
Discuss your requirements and refine your scope in a call with a Toptal domain expert.
2
Choose your talent
Get a short list of expertly matched talent within 24 hours to review, interview, and choose from.
3
Start your risk-free talent trial
Work with your chosen talent on a trial basis for up to two weeks. Pay only if you decide to hire them.
Top talent is in high demand.
Start hiring