2021-Ongoing
Working on telemetry investigations with explainable machine learning.
Working on NLP similarity and classification tasks.
Doing ML engineering for internal ML tools on Python, C#, Azure Cloud. Re-designed and coded a population unbiasing library in Python.
Mentoring and leading as tech lead in several projects.
2019-2020
Research computer vision and natural language understanding models applied to eCommerce scenarios
Natural Language Understanding models for chatbot intent and entity extraction. Developed a text similarity search system using feature vector indexes used for several projects.
2017
Demonstrate usage of herbaria images for Deep Learning Plant identification.
Creation of hierarchical loss functions for Deep Learning classification, particularly on plant taxa domain. Expertise on Theano, Caffe and Tensorflow
2014-2016
Develop debugging tools for GPU in C#, Python and C/C++.
2011-2014
Create business intelligence tools in SQL Server (SSIS and SSAS) and QlikView. Mostly ETL, dashboards and Data Cubes. Using Natural Language Processing and Pattern Detection techniques in C# to improve data quality. The new data tools resulted in several million dollars savings.
2008-2015
Developed software on Applied Statistics (.NET VSTO) www.statsolver.net
Created several Excel addins for companies such as DHL and Softland. Worked on a Cloud-Based Statistical Analysis tool with Spring MVC, Bootstrap+JQuery, Spring RESTful Web Services and MongoDB/PostgreSQL.
2008-2011
Support Call Center Transactional/Loyalty applications from legacy ones (C/C++/VB6.0) to new projects using .NET and Java.
Coding, design and integration with AS400, Tandem, Unisys, and SQL Server.
Developed skills in Softphone Genesys SDK and Microsoft CCF/Dynamics CRM/Sharepoint.
2003
Generated ArcView map layers from Costa Rica's plants/animals collection points, using Java and ORACLE Databases.
Developed basic experience with ESRI geographical tools (Arcview, Arcmap).