Uillian Luiz Ludwig

SAP Data Intelligence provides a set of operators so that customers can model ETL pipelines. In collaboration with my team, I developed operators to communicate with several databases and cloud storage, including MySQL, Oracle, IBM Db2, Snowflake, Amazon Redshift, Google BigQuery, and others.

The operators could read and write tables and views from databases, Parquet, CSV, Orc, JSON, and Excel files from cloud storage. Transformations included join, projection, union, and case operations and were processed by the SAP Data Services engine, which we embedded and orchestrated.

These operators were part of a pipeline sub-engine called Flowagent, which my team members and I also developed.

Ingestion jobs were one of the most used features of SAP Data Intelligence, and I added a feature to allow their parallel loading. This feature allowed users to define logical partitions in the pipeline UI. Then, in the back end, the engine orchestrated the execution of the jobs. It had a master/slave architecture, where the master oversaw delegating partition to the workers as well as any synchronization that was required between workers. Each worker was a Kubernetes Pod, and the communication between workers was done via HTTP.

Customers widely requested and used this feature because it provided faster processing time for their ingestion jobs. Besides, it didn't require native support from the source database, so it could be used with any database.

During my job at SAP, I developed a GitHub bot to help GitHub administrators with everyday tasks. The bot was created using Node.js, TypeScript, Express.js, and MongoDB.

Some of the features of the bot included:
• Pull requests merge queue managed via labels.
• Jira integration to automatically update and validate the state of backlogs and bugs.
• Test quarantine, which helped to overcome issues with flake tests.

Capacity, interference, and affinity-aware placement algorithm (CIAPA) is a set of algorithms for the smart placement of multi-tier applications on shared cloud environments. I developed this application as part of my master's program in computer science.

CIAPA can be divided into two parts:
• The application core contains models, equations, and optimization heuristics, including hill climbing and simulated annealing.
• The user interface, developed using Angular and TypeScript, allows easy interaction with the application core.

Node-tiers is an application that allows simulating several multi-tier applications types. It is useful for academic research, allowing to easily simulate different application workloads.

Its main features include:
• Several resource-intensive tiers were implemented, each with different characteristics.
• Network utilization in the communication with each tier is personalizable.
• The application can be executed with any tier configuration (1 to N-tiers).
• The addition of new application tiers is simple.
• It was built using Node.js, TypeScript, and MongoDB.

ÔnibusRS is a web and mobile system that locates the best routes for intercity buses of the Brazilian state Rio Grande do Sul.

The system was developed utilizing the following programming languages and technologies:
• Python, PHP, and MySQL on the back-end.
• JavaScript, jQuery, and Bootstrap on the front-end.
• Java and SQLite on the Android application.

The data was collected with the implementation of a web crawler in Java.

Mundo Brasfoot was a fan-based website for a soccer management game called Brasfoot. Users would share their content on the website, including extra teams, analysis of best clubs and formations, shirts and shields packs, and helper tools for the game.

Mundo Brasfoot was developed using WordPress, and it had an average of 3,000 unique views per day, while the helper tools had more than one million downloads. I created these tools using Visual Basic .NET, and it was my first contact with programming.