Arinti Case Study
Arinti is an Applied AI Solution Provider dedicated to helping organizations leverage the advantages of the AI revolution.
**CHALLENGE**
One of Arinti's key clients, a global fast-moving consumer goods (FMCG) company, needed to adopt DevOps within its data analytics environment and enhance its data lake architecture for the Benelux team to minimize waste in their product manufacturing. The project required the expertise of a Senior Data Engineer with extensive knowledge in Azure, Databricks, and Python, as well as advanced technical consulting abilities.
**SOLUTION**
Pluralit assigned Bruno Magri, a skilled data engineer from Brazil, to this project. He collaborated closely with the client to develop optimal solutions. To implement DevOps in the data analytics setting, Bruno consolidated various code versions into a final git branch, manually aligning and gathering correct code contributions from different team members. This effort connected the development (DEV), quality assurance (QA), and Production environments. He coordinated the development process with the project manager to ensure all DevOps pipelines were established with the appropriate continuous integration (CI) and continuous delivery (CD) processes across the client's environments. To enhance the data lake architecture, he introduced the Medallion Architecture, which consists of multiple data layers ensuring consistency, clear data lineage, and high data quality within the data lake house.
**RESULTS & BENEFITS**
The introduction of CI/CD in DevOps minimized human and data errors, enhanced product reliability, and improved efficiency and agility in delivering new features. The revamped data lake provided increased flexibility in managing data through a recognized design pattern, resulting in approximately €45,000 in annual cost savings. It also accelerated data delivery by reducing the number of tiers in the database and semantic layers stored in the cloud. These accomplishments enabled the end-client to boost profits, decrease waste, perform statistical analyses with all processed data, and plan production more effectively.