Iβm passionate about technology, software architecture, research, and innovation. Always looking for new challenges that keep my mind working and push the boundaries of my knowledge and experience, I have worked on many different projects. From pure backend with millions of RPMs to carefully designed performant frontends, and on monolithic apps, libraries, and microservices. Iβm interested in performance, code quality (Clean Code fanΒ π), scalable and clean architectural solutions.
Also, since 2018 I have been part of ORT Uruguay universityβs teaching staff. I have been in charge of teaching both theoretical and practical subjects. For example, in Programming 1, I give the students the first approach to programming from scratch by teaching the theoretical concepts and then I help them apply that to a particular language (Javascript); in Programming 2, the focus is on software design quality, and we see in more detail object-oriented programming (we see it in Java). And finally, Programming for Biotechnology is also the first approach to programming but targeted to a profile that needs more problem solving through programming and we focus rather on scripting. Having the opportunity to work at the same university where I got my degree has been a very rewarding and enriching experience and something I enjoy doing.
- Next js + React component library: Split a monolithic Rails app into 3 (2 new software pieces), a web app built in next js that would store all the new pages that will interact via RESTful APIs with the Rails backend and a component library built using typescript and styled-components to store all the generic/reusable React components.
- Marchine learning vectors: Purely backend project built using Java 8, Spark Java, Guice, Elastic search, and Google big table, on a microservices ecosystem.The main goal was to store machine learning data vectors for online and offline training and evaluation. The challenge here was managing the amount of data (millions of RPM and terabytes of information), and deliver the data for online evaluation (sync on a near-real-time basis) under 200ms of response time.
- Anomaly detection system: Monitor millions of payments a day trying to find anomalous and fraudulent trends (attacks). It was built using Java 8, Spark Java, Guice, and a fine-tuned MySql database.
- Chargebacks administration: Project built with Grails and Angular, to automate and manage chargebacks on a well-known payment platform in Latin America.
- Genexus: Code generation platform to make programs easier to build, with no programming skills required.