English:
Advanced
Mexico
UTC -06:00
America/Mexico_City
Highly skilled Data Scientist and Machine Learning Engineer with 6 years of experience in building predictive models for large-scale data sets.
- I developed and programmed a Siamese neural network model under my guidance, capable of predicting patient outcomes based on multiple factors. -I took the responsibility of maintaining legacy code, refactoring it as necessary and creating pertinent UnitTests. This dedication resulted in the update of more than 10,000 lines of code, maintaining a high standard for the codebase. - I innovatively implemented a Python solution combining pyTorch and OpenCV for the identification of theft events in supermarket stores. This initiative led to annual savings of over 3 billion USD for the client. - I optimized and systematized a computational pipeline for preprocessing and analyzing medical images, resulting in a remarkable 20-fold increase in output efficiency. -I improved existing ML algorithms, effectively reducing processing time requirements from 8 hours to just 10 minutes, while maintaining an accuracy rate of over 97%.
Want to hire this engineer?
Check if Diego is availableExpertise
Years of commercial development experience
6 years of experience
Core technologies
Other technologies
Project Highlights
The Home Depot
May `23 - Present
2 years
Work with a major store using different machine learning models to predict sales behavior depending on external events
Responsibilities & achievements
• Implemented Google Cloud Platform pipelines with Vertex AI. Acquiring data from BigQuery and deploying ML models on a Jenkins server. • Improved existing ML algorithms to reduce processing time requirements from 8 hours to 10 minutes and maintain an accuracy of over 97%. • Maintained legacy code as needed by refactoring and creating pertinent UnitTests to keep the code base to a high standard, resulting in over 10,000 lines of code being updated.
Corteza.ai
May `22 - May `23
1 year
Created an advanced Python implementation of pyTorch and opencv to aid with the identification of theft events in large supermarket stores
Responsibilities & achievements
• Created novel computer vision approaches to the identification of extremely similar items, leading to successfully classifying 90% of items. • Extracted, transformed, and loaded dozens of terabytes worth of unstructured data, allowing for the training of sophisticated deep-learning models. • Created an advanced Python implementation of pyTorch and opencv to aid with the identification of theft events in large supermarket stores, saving the client upwards of 3 billion USD annually.
Computational and Translational Neuropsychiatry Lab,
Aug `20 - Aug `22
2 years
Institute of Neurobiology
Optimize and systematize a computational pipeline for the preprocessing and analysis of medical images
Responsibilities & achievements
• Computed and processed over 1,000 human Magnetic Resonance Images to ensure their quality and relevance. • Developed, maintained, and analyzed a pre-clinical substance abuse disorder model that generated over 100 thousand data points. • Created analysis pipelines based on nilearn, a python-based tool for fMRI.
Cellular and Molecular Neurobiology Lab
Dec `17 - Apr `19
1 year
Institute of Biotechnology
To develop and program a Siamese neural network model capable of predicting patient outcomes according to multiple factors.
Responsibilities & achievements
• Optimized an RNA extraction technique, resulting in a 32% increase in extracted RNA quality and quantity. • Deployed a keypoint model based on pyTorch to analyze animal behavior in low-light conditions, saving the laboratory over 15,000 dollars in licensing fees. • Created models based on Tensorflow and pyTorch, with support of opencv, numpy, and pandas.
Education
Higher education in Computer Science
Agency
10-50
GMT-6
Ciudad Obregon/Mexico
Core Expertise
Industries
E-Commerce & Retail, Banking & Finance, Human Resources & Career, Safety & Security, Food & Drink, Arts & Culture
Want to hire this engineer?
Check if Diego is available