As a Freelancer:
Check out my profiles at Freelancer and Upwork
- I have worked with more than 20 clients from several countries, including Norway, India, Canada, Spain, Serbia, Saudi Arabia, Colombia, UK, Australia, and the USA, with excellent client satisfaction.
- Developed deep learning image classification systems including a clothes style classification model, an NSFW images detector, and a landmark detector in dental cephalometric x-ray images.
- Worked on academic and commercial NLP projects: text classification for healthcare and sentiment analysis for sociological research.
- Developed a sentiment detector in voice-mail messages.
- Build complete data science products, from exploratory data analysis to predictive models.
At Startups:
Machine Learning and Big Data Engineer at Makrwatch, from December 2017 to September 2018
I helped to made sense of data gathered from YouTube by creating Machine Learning models and Big Data analytic flows on cloud services, responding to specific client needs and helping to create the infrastructure for the company’s strategic vision
In Academia:
Research Assistant Instructor at Universidad Icesi (Cali, Colombia)
In 2015 I designed and taught an introductory Machine Learning course for graduate students, obtaining 6 over 7 points on the after-course student satisfaction survey.
From 2005 to 2011 I worked on several research projects in the areas of wireless network propagation prediction, network security, and high-performance computing, coauthored 11 papers published in national and international journals and conferences, and designed and taught an introductory Mobile Applications course during 8 semesters.
Ph.D. Student at Universidad de los Andes (Bogotá, Colombia)
I was a full-time student from 2011 to 2014 in system engineering, working in developing cloud computing for e-Science with QoS guarantees based on non-dedicated resources, and coauthored 3 papers published in conferences in Canada and the USA.
I also helped biology and chemistry research groups develop computationally intensive models supported by cluster and parallel computing.