Willian Hayashida

• Developed a Python framework for Power BI dashboards, enhancing productivity at scale. • Implemented CI/CD pipelines with Jenkins and GitHub Actions to automate testing and deployment. • Designed a scalable data architecture with end-to-end data lineage, job observability, centralized credentials and data source management, and usage metrics for product monitoring. • Led a team of 7 people, responsible for roadmap planning, architecture design, and driving creativity and collaboration within the team.

• Enhanced the offline Feature Store (PySpark) by implementing text processing and time-based aggregations. • Maintained the online Feature Store services (Kotlin) to support real-time user feature processing. • Developed and deployed 200+ features for User Income model, Loan Score and the App Marketplace with Open Finance data. • Conducted model and features backtesting in Databricks to evaluate and improve their performance.

•Built a central dashboard for supplier related defect costs used by 4000+ professionals worldwide. •Developed a Python data pipeline to collect and transform data for the central reports, reducing the data collection time by 95%. •Designed and developed the Central Supplier Quality Report for the Automotive Sector in Power BI, including a staging layer to model and transform the data with Pandas and MS SQL Server.

• Researched cloud infrastructure for HPC, automating cluster deployment on Azure and AWS (EC2, CloudFormation). • Optimized performance of BRAMS and FWI models through resource tuning, job tracing, and scenario analysis. • Published a paper in the 16th International Congress of the Brazilian Geophysical Society & Expogef doi. • Presented the analysis of cloud computing feasibility for distributed applications at WSCAD-WIC.