If you are like me, and you have devoted a good chunk of your professional career to the world of data analytics in the enterprise, I guess the following story will sound very familiar to you:

“You are tasked with a new mandate: — coming from the top management of the company — to transform the organization towards [insert most hyped buzzword at that time] lets say data-centric or data-driven. After a saga of countless meetings and long committees, great news! It is finally decided to invest a couple of million or so in software, hardware and consulting/implementation services.

Following…

I admit it, as a data engineer I have been full of envy and jealousy towards my colleagues — application engineers — in countless times.
When they were automating their production release and deployment generation processes, we had to waste hours drawing arrows and copying boxes from one ETL environment to another or manually resolving conflicts between different branches (if code versions on my colleagues’ laptops can be even called that) of a semantic model of a BI tool.

And is that any ETL or BI tool that comes to mind still has that 90s flavor — Why this obsession…

Lo admito, cómo ingeniero de datos he sentido envidia de mis compañeros — ingenieros de aplicaciones — en innumerables ocasiones

Cuando ellos estaban automatizando sus procesos de generación de releases y de deployments en producción, a nosotros nos tocaba perder horas copiando cajitas de un entorno de ETL a otro o resolviendo manualmente los conflictos entre diferentes branches (si es que se pueden llamar así a las versiones en los portátiles de mis compañeros) de un modelo semántico de una herramienta de BI.

Y es que cualquier herramienta ETL o BI que se me viene a la cabeza tiene aún…

The history of computing is a tale of increasing abstraction

No matter which technology you look at, the same pattern somehow repeats itself: a breakthrough happens in a research environment, public or private. At this stage just a very few people are able to understand this discovery and contribute to its refinement, this crowd tends to sign starting with PhD before his name. Expectations are high at this point.

After a viable business use of this technology emerges, one might expect a few companies trying to jump in and tame and understand this novel discovery. However the real widespread adoption only happens when the technology is democratised and “ordinary” IT…

Cuando nos enfrentamos al entrenamiento de una red neuronal, la decisión de que optimizador seleccionar parece estar envuelta en un halo de misterio, ya que en general la literatura alrededor de los optimizadores requiere de bastante bagaje matemático.

De cara a definir un criterio práctico, vamos a realizar una serie de experimentos para ver el comportamiento de diferentes optimizadores en problemas canónicos del aprendizaje automático. Así podremos elegir un optimizador de forma sencilla.

Marco teórico

En esencia, el objetivo del entrenamiento de redes neuronales es minimizar la función de coste encontrando los pesos adecuados para las aristas de la red (asegurando, eso…

Generating a working (value-generating) machine learning model is not an easy task. It usually involves advanced modelling techniques and teams with scarce skills. However, this is only the first step on an even more complex task: deploying the model into production and preventing its degradation.

Even being alleviated by the cloud shift, at least two-thirds of IT spent is still concentrated on maintenance-mode tasks. …