Yunior Cabrales

I am a Physicist with 10+ years of experience on statistics and biophysical system modeling. With my research I have contributed to fields like premature infant brain maturation and Alzheimer’s disease in the last 5 years. With 3+ years of experience on Data Analytics, I have a special interest in applying my skills and help companies to develop strategic plans based on predictive modeling and findings.

Magnetic Resonance Image Analyst.

Biomedical Res. and Innov. Inst. of Cadiz, Cadiz, Spain (INiBICA)                                          (Oct. 2018 to Date) 
- I have analyzed a 1000+ structural Magnetic Resonance images from premature infants using non-linear models and statistical analysis. Delineated white mater tracts on 200+ Diffusion Tensor Images (DTI) with open source tools like FSL and DTITK. Modified and adapted algorithms originally oriented to MR image analysis on adults to be used on premature, using both Bash and Python for an automatic performance. This approach leads to a 90% increase in productivity. I learned the principles of deep learning, convolutional neuronal networks (CNN) and their application to image processing.

Postdoctoral position.

Max Planck Inst. for Mult. Sciences, Goettingen, Germany                                      (June 2017 – Sept. 2018)
- I determined the structure of the 80% Tau/microtubule protein complex using Spectroscopy of Magnetic Resonance as well as specialized software for molecular dynamics simulation from 3 peptides of 30 amino-acid residues each. I found and improved methods to study in depth mechanisms of biological processes and published 2 articles in high-impact conferences and journals. I conducted 1 rotation lab of student from a master in Neuroscience at Georg August University. Taught “NMR principles” to master student in Organic Chemistry and on Neuroscience (50+ hours).

Data Science-related projects
- Fraud detection on credit cards. (link to project)
I Developed 5 predictive models that identifies a fraudulent bank transaction with a precision of 99% in a highly imbalanced dataset.
Used models: (Logistic Regression, KNN Classifier, Decision Tree Classifier, Random Forest Classifier, Ada Boost Classifier and Gradient Boosting Classifier)
- Hotel Booking cancellation prediction. (link to project)
I predicted hotel booking cancellation (acc. sore > 0.86) using either ensemble-based models or keras-based models. Exploratory data analysis, feature engineering, cross validation and Hyperparameter tuning

- PhD in Natural Sciences (2013-2017)
Georg August Univ., Göttingen, Germany.
Structural characterization of intrinsically disordered proteins by using NMR and Molecular Dynamics Simulation.

- Master degree in Biophysics (2011-2012)
Autonomous University of Madrid, Madrid, Spain .
Study of the polymerization Processes by using advanced fluorescence techniques.

- Bachelor in Physics
Universidad de Oriente University, Cuba.