Guillermo Barbadillo

Deep Learning Expert | Data Science Challenge Winner

Pamplona, Spain

Freelancer

Contact

ABOUT

Since 2014 I have been working on a startup as a Data Scientist. I’m expert on Deep Learning and I have worked on projects involving computer vision, face recognition and speech recognition among others.

During my free time I like taking part on Data Science Competitions. I have proven my skills winning DSTL's "Safe Passage" competition (27k$) where I had to create a model to detect and classify vehicles on satellite imagery. I have also achieved second position on Schneider's "Power Laws: Optimizing Demand-side Strategies" (8k$) where I had to optimize the use of a battery to reduce the electricity bill. I currently hold the rank of Master on Kaggle where my best ranking was 256 of 82.709 (0.3%). I have participated on more than 15 competitions covering all kind of data science fields: images, audio, text, tabular data...

I can provide world class solutions to your business.

LANGUAGES

English

Native or bilingual proficiency

Spanish

Native or bilingual proficiency

SKILL DETAILS

Technology

290 ILS / hr

UNIX, Linux

Ubuntu is my main OS.

Amazon EC2 - Elastic Compute Cloud

I have used EC2 for training and deploying models.

Docker

I have used it many times, also nvidia-docker.

AI, Artificial Intelligence

Artificial Intelligence is my passion.

Robotics

I did two internships in robotics and I build my own sumo robot during my master.

Machine Learning

I have participated in many data science competitions were I have used all kind of machine learning algorithms.

Computer Vision

Most of my work has been in the field of computer vision. I'm an expert on applying deep learning to classify or segment images.

Neural Networks

Deep Neural Networks or Deep Learning is the state of the art on artificial intelligence.

Speech, Voice Recognition

I won silver medal on Google's Tensorflow Speech Recognition Challenge.

Deep Learning

I'm a Deep Learning Expert. I have applied it to many different kind of data: images, audio, text, tabular data...

Software & Programming

290 ILS / hr

Python

Expert level

Data Scraping

Gathering data is essential to train good models.

EMPLOYMENT HISTORY

November 2012 - Today

Data Scientist

das-Nano

I have developed Deep Learning systems for different applications.

EDUCATION

2007 - 2012

Master in Electronics and Automation Engineering

UPNA

PATENTS & INVENTIONS

May 2018

Counting stacked planar substrates

US20180121782A1

Devices and methods for counting planar substrates stacked on a first plane are disclosed. Image capturing sensors are arranged collinearly in a manner substantially perpendicular to the first plane. The image capturing sensors are configured to capture images of portions of counting sides of the stacked planar substrates. The image capturing sensors are configured so that the images captured by every two consecutive image capturing sensors comprise an overlapping portion of the stacked planar substrates.
See patent Counting stacked planar substrates

May 2017

3d printing material encoding

US20170120528A1

3D printing material blends (110) are disclosed configured to be used in a 3D printer. The 3D printing material blends comprise a 3D printing material portion and a magnetic nanoparticle portion (S1, S2, S3, S4). The magnetic nanoparticle portion is embedded in the 3D printing material portion in such a way that the 3D printing material blend is encoded with a uniquely identifiable predetermined code. Furthermore, 3D printing units and systems having sensors for identifying 3D printing material blends and methods of encoding 3D printing material blends are also disclosed.

September 2014

MEASURING MAGNETIC SECURITY FEATURES ON SECURITY DOCUMENT CONFIGURATIONS

PCT/EP2014/061844

Apparatuses and methods for measuring magnetic properties of magnetic security feature are disclosed. Magnetic reader modules are movably coupled to support modules configured to guide the magnetic reader modules along a first direction. The magnetic reader modules are configured to perform contactless measurements of magnetic property changes induced as the magnetic reader modules move above said magnetic security features and generate an electrical signal in response to said measurements.