I’ve worked as programmer and currently project leader of LOVE, the LSST Operators Visualization Environment, a system to monitor several components of the Vera C. Rubin Observatory. This Observatory has over 65 subsystems that monitors different aspects of the telescope and its environment. LOVE is connected to the mayority of these systems.
I started working in LOVE when it had 1 year of development. At the beginning I performed as programmer having to understand the basis of a complex and beautiful system. After some time I changed my role to proyect leader, shortening my time programing but also giving full time to the needs of the project in its entire spectrum.
What I’ve done here:
Management and planification of the project schedule.
Support in UI/UX decisions.
Define sprint objectives and lead a team of 4 persons.
Develop ReactJS features and components.
Work on deployment infrastructure: Docker & Kubernetes.
Documentation for different components of the system.
Adding testing for frontend and backend features.
Support requirement verification and validation process.
Design methods to ensure the system’s performance.
I also had the opportunity to visit the Observatory, fulfilling one of my dreams. You can check me working on the right side of the following picture:
This project is Open Source so I invite you tu check the repositories:
I’ve worked implementing a lot of LMS systems (Learning Management Systems) used for e-learning education. I’ve developed several components of different kind, in Frontend and Backend. Also worked with a lot of clients, using and agile approach with regular meetings to indentify, discuss and validate requirements.
How
Frontend: HMTL, CSS, CSS3, SASS, Javascript
Backend: AMP stack (Apache, Mysql -Maria DB- and PHP)
I’ve participated on the creation an entrepreneurship company called OSARE, we focused on developing innovative apps. This was one of my first job approachs, so it was the beginning of my development carrer. We were able to develop two applicaitons:
HelpyCar: an application used to geolocate car workshops in case you required any urgent service. This was developed in native Android and we never did a release, but it served as a first approach to the software development.
Dubbin: an entertainment application used to create dubs of scenes of any video you uploaded to the app. I’ve participated on the firt release of the app which was developed using native Android. I didn’t participate on the last release which got published on the Apps Stores.
How
Frontend: Android Native
Backend: AMP stack (Apache, Mysql -Maria DB- and PHP)
The discipline of astroinformatics has grown a lot over the past few years thanks to the creation of bigger and more sophisticated telescopes, such as the Atacama Large Millimeter/submillimeter Array. With better spectral resolution in data, a new challenge is set in the way astronomical data is analyzed. In particular, data cubes produced by radioastronomy projects have generated an explosion in the volume of data retrieved. Some tasks, such as
the identification of spectral lines becomes more complex. For this reason it is essential to develope accurate analysis tools that allows data to be processed automatically.
This works propose a novel method in the way spectra can be classified. The approach is based on an algorithm used in the world of Text Mining, named Latent Dirichlet Allocation, a probabilistic generative model capable of describing documents as a random mixture of words over topics. Here, each spectrum is represented as a mixture of transitions over species. A spectral line transitions database named Splatalogue is used to train different models based on the type of observed object or ALMA band. The algorithm is evaluated using the model to analyze real world data cubes and spectral line surveys from radioastronomy observations of ALMA. The main advantage of the proposal is the ability to model sparse and high dimensional data using posterior inference to classify new spectral observations. Results shows that L-LDA can be used to clasifiy spectral lines on data cubes with up to 97 % of accuracy.
