Team leader of different AI projects for the entity in charge of managing the public health system of the province of Quebec (RAMQ).Completed projects:- Assessment of the risk of non-conformity of medical billing without related clinical activity.- Automated Medical Request Acceptation for the Exception Drug Program.- Detection of potentially inappropriate access to Dossier Santé Québec (DSQ).- Assessment of the risk of inadmissibility to the Quebec public health system.Ongoing projects:- Evaluation of the risk of inadmissibility to the public drug insurance program.- Interpretable financial prediction of the consumption of medical and pharmaceutical services in the Quebec public health system.

I have been directing and developing the integration of machine and deep learning tools with different work teams. The main projects have been focused on flood early warning systems, optimization frameworks, and probabilistic forecasting. I worked closely with City Mayor of different cities in Colombia, frequently subsequent to flood disasters.Successes included integrating radar and satellite information with machine and deep learning models to define real-time operational systems. In one of the developed systems, a first alert was already launched by the system saving the lives of hundreds of people.

After one of the worst socioeconomic disasters in Colombia due to the 2011 floods, I led the coupling of a streamflow forecasting model with deep learning, data from the NOAA and the ECMWF, and twelve hydrological models in La Mojana region.We achieved it with a fraction equivalent to approximately 30% of what was normally invested in this kind of projects, established the bases of a high reliability Flood Early Warning System with a satisfactory lead time of 72 hours.

I developed the coupling of hydrometeorological models with post-processing of data with machine learning optimization techniques.The challenges of this project focused on the internal comprehension of high complexity models, the Python coupling of models programmed in Fortran and C ++, and the optimization of the linkage with evolutionary algorithms, Bayesian optimization and the DDS.The results of this project opened the way for the province of Quebec to explore a new stage of development of its operational hydrological model.

In charge of the course for graduate students "Analysis and modelling of environmental series" - given in French. The final work of the course, focused on a parallel evaluation of different neural network models by groups: perceptron with evolutive training, recurrent networks, and echo state networks, which was exhibited at the European Geosciences Union Conference.

I designed the hydrometeorological monitoring and Flood Early Warning System in three micro-basins with a community participation program in Caldas. I also worked as a consultant for SIE in the climatology, hydrology and hydroelectric potential of the Dulce River and the Quebrada Negra.

I participated in several projects related to the modelling of urban watersheds in the search for the optimization of existing infrastructure or the design of new civil works for flood control.

I led the evaluation of the supply, distribution, sanitary and pluvial water systems of 13 cities with more than 1 million people in the lower basin of the Bogota River. This work served the Regional Corporation of Cundinamarca as support for the improvement plan to be financed by the Inter-American Development Bank.

I led the design and construction of a 10000-square-foot ecological medical center. The construction was carried out with native woods (Guadua and Arboloco) and articulated metal joints.

I was in charge of the modelling of the water distribution and sewerage systems of the city of Tocancipa. I also worked on the optimization of the distribution system with genetic algorithms.The main achievements in the design focused on the optimization of a pumping system that supplied around 4,000 people and the location of two new supply tanks in the city to reduce the high pressures of the system.