Luiz Verçosa, Ph.D.

Senior Data Scientist @ bp

About

With a 𝗣𝗵𝗗 𝗶𝗻 𝗖𝗼𝗺𝗽𝘂𝘁𝗲𝗿 𝗘𝗻𝗴𝗶𝗻𝗲𝗲𝗿𝗶𝗻𝗴 from 𝗨𝗻𝗶𝘃𝗲𝗿𝘀𝗶𝗱𝗮𝗱𝗲 𝗱𝗲 𝗣𝗲𝗿𝗻𝗮𝗺𝗯𝘂𝗰𝗼 and an exchange at 𝗥𝗪𝗧𝗛 𝗔𝗮𝗰𝗵𝗲𝗻 𝗨𝗻𝗶𝘃𝗲𝗿𝘀𝗶𝘁𝘆, I bring expertise in Machine Learning and Data Science. Beyond my professional life, I'm a 𝗰𝗼𝗹𝗹𝗮𝗯𝗼𝗿𝗮𝘁𝗶𝘃𝗲 𝗮𝗻𝗱 𝗵𝗮𝗿𝗱𝘄𝗼𝗿𝗸𝗶𝗻𝗴 individual passionate about puzzles, chess, and sports in general. My international experiences include 𝗮 𝘆𝗲𝗮𝗿 𝗶𝗻 𝗖𝗮𝗻𝗮𝗱𝗮 𝗮𝘁 𝘁𝗵𝗲 𝗨𝗻𝗶𝘃𝗲𝗿𝘀𝗶𝘁𝘆 𝗼𝗳 𝗧𝗼𝗿𝗼𝗻𝘁𝗼 during undergraduate studies in Computer Engineering and 𝘀𝗶𝘅 𝗺𝗼𝗻𝘁𝗵𝘀 𝗮𝘁 𝗥𝗪𝗧𝗛 𝗔𝗮𝗰𝗵𝗲𝗻 𝗨𝗻𝗶𝘃𝗲𝗿𝘀𝗶𝘁𝘆 𝗶𝗻 𝗚𝗲𝗿𝗺𝗮𝗻𝘆 during PhD in Applied Machine Learning. I am fluent in English and intermediate in Spanish. As a data scientist with 𝗼𝘃𝗲𝗿 𝘀𝗶𝘅 𝘆𝗲𝗮𝗿𝘀 𝗼𝗳 𝗲𝘅𝗽𝗲𝗿𝗶𝗲𝗻𝗰𝗲, I have applied Machine Learning to solve real-world problems such as predictive maintenance, credit risk assessment, and generative AI for automation. 𝗛𝗶𝗴𝗵𝗹𝗶𝗴𝗵𝘁𝗲𝗱 𝗔𝗰𝗵𝗶𝗲𝘃𝗲𝗺𝗲𝗻𝘁𝘀: - Led predictive maintenance solution at Universidade de Pernambuco with Ball, using ML to forecast equipment failures from sensor data. - Developed LLM-based agent at e-Core to automate web scraping Python scripts, reducing time from 40 to 2.5 days and costs from $3,600 to $800 per 200 sites. - Built scalable ETL pipeline at Neurotech to support ML products, including credit risk and vehicle theft risk evaluation. - Published over 15 scientific papers in international journals and conferences (https://bit.ly/vercosa_publications). 𝗖𝗼𝗿𝗲 𝗦𝗸𝗶𝗹𝗹𝘀: - 𝗣𝗿𝗼𝗴𝗿𝗮𝗺𝗺𝗶𝗻𝗴 & 𝗟𝗶𝗯𝗿𝗮𝗿𝗶𝗲𝘀: Python, Java, Scikit-learn, XGBoost, LightGBM, Pandas, NumPy, SciPy - 𝗗𝗲𝗲𝗽 𝗟𝗲𝗮𝗿𝗻𝗶𝗻𝗴 & 𝗙𝗿𝗮𝗺𝗲𝘄𝗼𝗿𝗸𝘀: TensorFlow, PyTorch, LSTM, GRU, Transformers - 𝗖𝗼𝗺𝗽𝘂𝘁𝗲𝗿 𝗩𝗶𝘀𝗶𝗼𝗻: Fast R-CNN, YOLO, OpenCV, Tesseract, Ultralytics - 𝗚𝗲𝗻𝗲𝗿𝗮𝘁𝗶𝘃𝗲 𝗔𝗜 & 𝗟𝗟𝗠𝘀: Amazon Bedrock, Hugging Face, SageMaker JumpStart, RAG - 𝗕𝗶𝗴 𝗗𝗮𝘁𝗮, 𝗖𝗹𝗼𝘂𝗱 & 𝗜𝗻𝗳𝗿𝗮𝘀𝘁𝗿𝘂𝗰𝘁𝘂𝗿𝗲: AWS (S3, EMR, Glue, Hive, Lambda, Step Functions, EC2, SageMaker, Trino), PySpark, SQL, PostgreSQL, Docker, Git, Matplotlib, Seaborn - Other Tools: ARIMA, DBSCAN, K-means, Agglomerative Clustering, Process Mining, Django, REST API, React 𝗖𝗲𝗿𝘁𝗶𝗳𝗶𝗰𝗮𝘁𝗶𝗼𝗻𝘀: - AWS Certified AI Practitioner - AWS Certified Machine Learning Engineer – Associate Let's connect to discuss Data Science, AI trends, and market innovations. I'm open to sharing experiences, delivering talks, and consulting. Send me a message!

Country

Brazil

City

Recife

Industry

Computer Software

Skill

Generative AI, Microsoft Azure, YOLO, Long Short-term Memory (LSTM), Amazon S3, trino, Hive, docker, Amazon Elastic MapReduce (EMR), AWS SageMaker, Amazon Bedrock, Scikit-Learn, PyTorch, AWS Step Functions, AWS Lambda, Deep Learning, Large Language Models (LLM), Applied Machine Learning, Git, Problem Solving

Experience

Senior Data Scientist

2025-10 - Present · 1 yr

London Area, United Kingdom

Senior Data Scientist

CESAR

2025-3 - 2025-10 · 8 mos

Recife, Pernambuco, Brazil

As a Senior Data Scientist on a 10-person cross-functional team (including data scientists, DevOps, designers, and domain experts), I led the restructuring of the project repository in my first three months at CESAR, implementing best coding practices and enabling efficient local development and testing using Docker, Trino, and related tools. Recently, I’ve focused on predictive maintenance, conducting exploratory data analysis and modeling with industrial sensor data to forecast equipment failures and optimize operations.

Data Scientist

e-Core

2024-6 - 2025-3 · 10 mos

Rio de Janeiro, Brazil

At e-Core, I worked as a Data Scientist either alone or in small 1-3 person teams on agile consulting projects. I led both client interactions as well as the design of AI solutions for distinct business purposes. A worth highlighting project was the development of an LLM-based agent to generate and correct web scraping Python scripts at scale using AWS services. This solution had a major business impact, reducing development time for such scripts from ~40 days to ~2.5 days and cutting costs from $3,600 to $800 per 200-site expansion. Other projects included creating ML models for predictive maintenance to forecast machine failures from telemetry data, developing forecasts for container arrivals, and building models to predict vehicle time-to-sale.

Data Scientist

Ball Corporation

2022-2 - 2024-4 · 2 yrs 3 mos

In a 2-person R&D team (myself and one developer, supported by two managers) at Ball Corporation, I served as Data Scientist, developing an ensemble ML model for predictive maintenance on industrial machines. My role involved forecasting critical sensor values to enable proactive maintenance alerts, optimizing operational planning. Outcomes included the reduction on unplanned downtime by 23%.

Data Scientist

Pumatronix

2020-6 - 2022-1 · 1 yr 8 mos

Collaborating in a 3-person team (myself and two engineers in software and vision domains) at Pumatronix, I served as Data Scientist, developing computer vision models for license plate recognition and building the backend for a web system enabling human-in-the-loop interaction to correct model predictions on license plate characters. My role focused on character localization for OCR processing and deelopment of Django backend, ensuring accurate detection and seamless integration. Outcomes included delivering a web system with integrated models.

Data Scientist

Neurotech | Uma empresa B3

2017-4 - 2020-5 · 3 yrs 2 mos

Recife Area, Brazil

On a 6-person analytics team (including data scientists, engineers, and risk analysts) at Neurotech, I served as Data Scientist, establishing scalable data infrastructure for credit risk assessment and related products. My strategic role included automating ETL pipelines, conducting feature engineering and correlation analysis, and building predictive models. Outcomes included optimized data collection for multiple applications (e.g., credit and vehicle theft risk) reducing processing time by 30-40%, and delivering a durable system still in use today, enhancing lending decisions and operational efficiency.