Jelther – Apache Airflow, Microsoft Azure, Apache Spark

Company operates across Latin America and is one of the largest online commerce ecosystems in the region and facilitates transactions between buyers and sellers, covering various categories such as electronics, fashion, home goods, and beyond.

Jelther carried out such responsibilities:

• Led documentation refactoring to improve user experience towards feature engineering work;
• Mentored and developed a team responsible for a Machine Learning feature store platform company-wide that would allow building, testing and publishing their features for machine learning models;
• Conducted blocker solutions, reviews and assignment of tasks among the team members;
• PR reviews of new features to improve accuracy and costs of a custom feature store platform.

This platform is part of one of the largest food delivery companies in Europe that connects users with a wide selection of restaurants in their area, allowing them to order meals for delivery or pickup with ease. It enables customers to browse menus, place orders, and track deliveries in real-time, all from the convenience of their mobile devices or computers.

Among others, Jelther was in charge of the following:

• Maintenance of data orchestration tools based on Apache Airflow, used by 1000+ users;
• Built the transformation tools for the Data Teams;
• Data support on the usage of transformation tools;
• Worked closely with data scientists and analysts to provide the best approach and minimize costs;
• Implemented DBT orchestration tools on Apache Airflow to expand usage by other Data teams.

This platform is a global leader in fraud protection solutions, primarily focused on e-commerce businesses. Leveraging advanced technology and industry expertise, it offers comprehensive fraud prevention strategies and real-time monitoring solutions to ensure secure and seamless transactions for both merchants and consumers.

Jelther has successfully performed the following tasks:

• Led junior data engineers on data engineering projects to maintain transactional feature store databases and an analytical database used by data scientists to build and improve existing machine learning models;
• Built a custom ETL tooling to provide low latency data for machine learning models for fraud detection;
• Proposed architectural solutions for the enhancement of machine learning and fraud detection systems.