Developing an LSST alert stream broker: Fink (PI together with J. Peloton and E. Ishida).
Constructing a recommendation system for follow-up of supernovae for cosmology RESSPECT.
Photometric classification and machine learning.
In 2019, I released the framework SuperNNova , an open-source photometric supernova classifier that uses Deep Learning's Recurrent Neural Networks and Bayesian RNNs. We make emphasis on reproducibility, the code is in GitHub and statistical robustness of our methods (in particular uncertainty interpretability). SuperNNova is being applied to a variety of science cases including the DES 5-year photometric sample and Fink broker.
