Machine Learning Researcher with NLP/CV Focus

Heriberto A. Carbia

Personal Data

Place and Date of Birth: San Juan, Puerto Rico 28 december 1977 Address: Poppy St. A-26, San Juan, PR, USA

Phone: +1-305-***-****

email: *******@*******.***

linkedin: http://www.linkedin.com/in/heriberto-a-carbia-553a71a9 Work/Research Experience

2022-2024 Research Assistant at Riccardo Papa Lab, at University of Puerto Rico, Rio Piedras

Continue research on machine learning pipeline to predict the wing-color patterns of He- liconious erato butterflies from their genomic data using a convolutional neural network

(CNN) and a temporal neural network (TCN). My task included performing different types of error analyses as first steps for the preparation of a manuscript. 2021 NASA Machine Learning intern at Goddard Space Flight Center (GSFC), The internship consisted on working on the Concurrent Artificially-intelligent Spec- trometry and Adaptive Lidar System (CASALS) project under the mentorship of NASA engineer James MacKinnon. The project consists on developing a polar-orbiting Small- Sat observing system which integrates adaptive lidar, hyperspectral imaging and on- board artificial intelligence (AI) technologies. My task was to develop a machine learn- ing pipeline using LIDAR spectrometry data from the Earth’s surface to predict canopy cover and forest type using a temporal convolutional network (TCN). The overall ob- jective is to be able to address five Earth Science Survey observable recommendations: ecosystem structure, ice elevation, snow depth and water equivalent, topography and 3-D vegetation and the atmosphere boundary layer.

2018-2021 Research Assistant at Riccardo Papa Lab, at University of Puerto Rico, Rio Piedras

The research consisted of developing a machine learning pipeline to predict the wing- color patterns of Heliconious erato butterflies from their genomic data using different types of neural networks (RNN, LSTM, CNN, Transformer Networks). 2004-2014 Architectural Illustrator/3D Modeler at HJC Architects, San Juan Residential and Commercial Architecture

In charge of 3D modeling/rendering and presentations of both residential and commercial projects using vector-based design and modeling software including AutoCAD, 3ds max, and Sketch-up.

Education

August 2019-May 2021 Ph.D student in Computational Mathematics and Statistics, at University of Puerto Rico, Rio Piedras.

Courses: Real Analysis, Complex Analysis, Optimization, Modern Algebra, Graph Theory, Independent study on Machine Learning with Remi Megret. CGpa: 3.30/4.00

August 2017-May 2019 Master of Science in Applied Mathematics, at University of Puerto Rico, Rio Piedras.

Courses: Computational Analysis, Probability, Data Structures, Algorithms, Statistics, Linear Models

CGpa: 3.87/4.00

Education

August 2015- May 2017 Undergraduate courses in Mathematics, at University of Puerto Rico, Rio Piedras.

Courses: Calculus II-III, Advanced Calculus, Probability, Statistics, Linear Algebra, Abstract Algebra, Discrete Mathematics, Linear Programming, Numerical Analysis, Ordinary Differential Equations, Programming

CGpa: 3.71/4.00

May 2000 Bachelor of Science in General Sciences,

at University of Puerto Rico, Rio Piedras.

Major: Biology

CGpa: 3.13/4.00

Fellowships and Awards

May 2016 Dean’s List

Aug 2017 Claude-Shannon Award

from the Scholarship Fund for Excellence in Computer Science and Mathematics,

sponsored by the National Science Foundation (NSF) Aug 2019-2020 NASA PR Space Grant Fellowship

Aug 2020-2021 NASA PR Space Grant Fellowship

Certificates and Exams

June 2017 FM Actuarial Exam (SOA)

Nov 2021 Machine Learning - Stanford University (A. Ng) verifiy at http://www.coursera.org/verify/B3Z5LNFQCBFM Dec 2021 Neural Networks and Deep Learning - DeepLearning.AI (A. Ng) verifiy at http://www.coursera.org/verify/B6GJ6SWUXE2Y Currently enrolled Deep learning Specialization (5 courses by deeplearning.ai) Presentations

March 2019 Convolutional deep learning to predict physical characteristics from genomic data.

SIDIM 2019, University of Puerto Rico - Humacao.

April 2019 Genotype-to-phenotype: A deep learning approach. 3rd Annual Drosophila Meeting, Neurobiological Institute - San Juan Oct 2019 Predicting phenotypic traits from genomic data using deep learning. NASA Technology Road Tour - UPR-RP, San juan

Nov 2019 Predicting phenotypic traits from genomic data using deep learning. Neurobiology Seminar, CRIAS - UPR-RP - San Juan

Manuscripts

Published Perfect mimicry between Heliconius butterflies is constrained by genetics and development. Journal: Proceedings of the Royal Society B.

Authors; S.M. Van Belleghem, P.A. Alicea Roman, H. Carbia Gutierrez, B. Counterman, and R. Papa.

In preparation Genotype to phenotype: A convolutional deep learning approach. Co-authors: Steven Van Belleghem, Remi Megret, Edusmildo Orozco, and Riccardo Papa.

Computer and Math Skills

Programming/ML: Artificial Intelligence, Machine Learning, Artificial Neural Networks, Deep Learning, TensorFlow, Keras, Data Science, python, R, RStudio, MatLab, C++, mathematica, Pandas, Scikit-learn, NumPy, MatPlotLib, Clustering (K-means), PCA, Regression (linear and logistic), Segmentation, Research

Statistics/Probabilty: Regression, Hypothesis Testing, Probability and Inference, Predicitive Modeling

Architecture/Modeling: AutoCAD, 3ds Max, Photoshop, Illustrator Microsoft: Excel, Word, PowerPoint

Languages

Spanish: Fluent

English: Fluent

Interests and Activities

Artificial Intelligence, Deep Learning, Machine Learning, Data Science, Mathematics, Actuarial Sciences, Scuba duving, Football, Tennis, Travelling.

Contact this candidate