Victor Blaga

A Java-based stream processor that enforced business rules validations on status messages generated by the assets of the client. I was the tech lead of a team of four developers that implemented a rule engine capable of validating business rules on high-throughput streaming data. Our main challenges were stateful stream processing under a high load (more than 50 messages per second) and also achieving and maintaining a high team velocity, as we implemented close to 100 business rules in a relatively short time.

Because of sound architectural decisions and solid development practices, the project was a great success and the client was very happy with our team's performance.

A mostly Python (and some Ruby on Rails) project, hosted on AWS, to scrape hotel reviews from various review websites (e.g., TripAdvisor, Booking.com, Expedia, and so on), and run them through machine-learning-based natural language processing algorithms.

The goal was to extract topic-based scores from the textual reviews and aggregate them into a global hotel category score, e.g., a review such as "the hotel was close to the city center" would score "good" on the "location" topic and would contribute positively to the global hotel location score.

I was the lead engineer, responsible for implementing the scrapers, the analysis modules, and the infrastructure components around them. I also had to coordinate with various project stakeholders and contributors, including data scientists - to integrate the ML models and with product owners - to calibrate scoring aggregation methods.

I also designed and developed a couple of tools around the main project: an experimentation tool used by data scientists to iterate and fine-tune models and a classification tool used by language experts to generate training data.

A Java-based REST service—hosted on OpenShift and Kubernetes—to automatize and reduce to a minimum the manual labor involved in the goods-receipt process at manufacturing plants.

I coordinated with business stakeholders to refine and implement business requirements and with SAP specialists to define data selection processes. I was the primary (and only) developer on this project.

A Ruby on Rail server-side rendering website, where I was the main back-end developer and coordinated with the designer to build the HTML views. I was also responsible for the front-end JavaScript development—implemented with jQuery—providing features such as automatic-refresh, infinite scrolling, and so others.

Additionally, I build web scrapers capable of retrieving social media reviews for various hotel and restaurant websites as well as a machine learning-based natural language processing module capable of analyzing textual reviews and deriving categorical scores.