#Personal – “Cindy Healy, always check the cables” or the best way to encourage you to study an engineering career #GoalCast #GoalCast


There are many programs to promote diversity in work environments. Many of these programs also focus on encouraging women to study engineering careers.

In my particular case, I think that the next 4 minutes are the best way to explain what you can do by being an engineer.

Important: If you’ve read the book, and you’ve seen the movie, what Cindy Healy says it’s becomes much more important!

Greetings @ Toronto

El Bruno



#Personal – “Cindy Healy, siempre comprueba los cables”, o la mejor forma de animarte a estudiar una carrera de ingeniería #GoalCast


Hay muchos programas para promover la diversidad en los ambientes de trabajo. Muchas de estos programas también se centran en animar a que las mujeres estudien carreras de ingeniería.

En mi caso particular, creo que los siguientes 4 minutos son la mejor forma de explicar lo que puedes hacer siendo un ingeniero o ingeniera.

Importante: Si has leído el libro, y has visto la película, ¡lo que comenta Cindy Healy tiene el doble de valor!

Saludos @ Toronto

El Bruno


#AI – Open Neural Network Exchange, Facebook and Microsoft help us to change between different AI Frameworks #ONNX



When a platform or technology begins to be popular, it often happens that Frameworks supporting this technology begin to appear as mushrooms in a wet forest in spring. JavaScript may be the best example of this.

Well, the same thing is happening in the world of artificial intelligence. The most common tools I found that specialists uses are Caffe2, CNTK, TensorFlow, and others. However, the interoperability or change between these tools is a problem with not an easy solution.

Well, 2 big players like Facebook and Microsoft, have agreed to help with the migration and interaction between these tools creating an interesting project ONNX.

ONNX is a community project created by Facebook and Microsoft. We believe there is a need for greater interoperability in the AI tools community. Many people are working on great tools, but developers are often locked in to one framework or ecosystem. ONNX is the first step in enabling more of these tools to work together by allowing them to share models. Our goal is to make it possible for developers to use the right combinations of tools for their project. We want everyone to be able to take AI from research to reality as quickly as possible without artificial friction from tool chains.

ONNX allows models to be trained in one framework and then transferred to another. The models currently compatible with Caffe2, CNTK, MXNet and PyTorch. In addition, there are connectors for many other frameworks.

Looking at the GitHub repository, it really impressed me that import and export scenarios from TensorFlow (Google) are included in the product roadmap.


Happy Coding!

Greetings @ Calgary

El Bruno



#AI – Open Neural Network Exchange, gracias a Facebook y a Microsoft ahora podemos utilizar e interactuar con diferentes AI Frameworks #ONNX



Cuando una plataforma o tecnología comienza a ser popular, suele suceder que los Frameworks de soporte a esta tecnología comienzan a aparecer como setas en un bosque húmedo en primavera. JavaScript puede ser el mejor ejemplo de esto.

Pues bien, lo mismo esta sucediendo en el mundo de la inteligencia artificial. Las herramientas mas comunes suelen ser Caffe 2, CNTK o Tensorflow. Sin embargo, la interoperabilidad o cambio entre estas herramientas es un problema de no fácil solución. Pues bien, 2 grandes de la industria como son Facebook y Microsoft, se han puesto de acuerdo para ayudar con la migración e interacción entre estas herramientas. Y es de este acuerdo, que ha surgido ONNX.

ONNX es un proyecto comunitario creado por Facebook y Microsoft. Creemos que es necesaria una mayor interoperabilidad en la comunidad de herramientas de IA. Muchas personas están trabajando en grandes herramientas, pero los desarrolladores a menudo están encerrados en un marco o ecosistema.

ONNX es el primer paso para permitir que más de estas herramientas trabajen juntas al permitirles compartir modelos. Nuestro objetivo es hacer posible que los desarrolladores utilicen las combinaciones correctas de herramientas para su proyecto. Queremos que todos puedan llevar la IA de la investigación a la realidad lo más rápido posible sin la fricción artificial de las cadenas de herramientas.

ONNX permite que los modelos sean entrenados en un framework y luego transferidos a otro. Los modelos actualmente compatibles con Caffe2, CNTK, MXNet y PyTorch. Además, existen conectores para muchos otros frameworks.

Dando un vistazo a su repositorio de GitHub, llama la atención que en el roadmap del producto se incluyen escenarios de importación y exportación desde TensorFlow (Google).


Ahora que tengo la oportunidad de “mover modelos” pues me ahorro la necesidad de conocer a fondo los nuevos.

Happy Coding!

Saludos @ Calgary

El Bruno



#Opinion – News and more news on Artificial Intelligence and we can expect in the near future


At the beginning of this year, I participated in a meeting with the general manager of Avanade Canada at the Microsoft Technology Center, where we presented our global Vision on technologies and trends for this year. It’s named “Avanade Techvision 2017” and this year is the first time that all the topics we talked about were covered under a general theme: Artificial Intelligence.


Over the next 5/10 years evolution in AI will impact on the way the society works. Obviously here we need to talk about topics such as creating new jobs, augmenting and optimizing existing jobs with devices like Hololens, and many other changes. In example, in the near future, thanks to artificial intelligence, we can detect signs of diabetes in our body, only using the camera of our smartphone. A photo analyzed in the device, supported by a DNN, can quickly tell us if we are at risk for diabetes. (See references)

However, an interesting detail in this aspect is that the current hardware needs to upgrade at lot to get there. Nowadays, the consumer focused hardware is the one that has to accelerate to be able to live up to it. In one hand we have big players, like specialized laboratories with enough money, and they already have special computers which give them the ability to apply AI algorithms. The big challenge is like bringing the IA to the final consumer in an affordable way. Without large data exchange costs using the cloud, and with processors which are battery efficient.

Yesterday I wrote about the new HPU processor with AI capabilities that will be incorporated in the new version of Hololens. Microsoft has already been responsible for developing the V1 of this HPU processor, and it is no wonder that this new processor, which some call “AI co-processor”, not only has applications in Hololens, but also in other devices such as smartphones, laptops and tablets.

Today I read in Wired the article “The rise of AI is forcing Google and Microsoft to become Chipmakers“, which describes the approach that Google also has about it. HPU incorporates the concept of a re-programmable chip, where for example for Hololens you can add AI capabilities for voice recognition and recognition of movements and gestures with the hands (naturally necessary in Hololens).

Another example is Google. Google works on a processor called Tensor Processor Unit. This processor also implements DNN capabilities, and has saved Google the creation of 15 Data Centers for speech-recognition-related activities. I guess, that 15 Data Centers out of the house improvement budget, should be an interesting savings for the board of directors of Google.

Note: TPU is nominated based on TensorFlow. TensorFlow is an AI system originally created by the Google Brain team and was published in open source mode a few years ago. It’s the core of a lot of Google’s internal work and is widely supported by the AI developer community.


And if this were not enough, the CEOs of a couple of big companies start to have some not nice words between each others on the Artificial Intelligence world. In this case it is Elon Musk (aka Tony Stark) who in tweet said that the knowledge of Mark Zuckerberg on AI is limited.

In reality, this arises because they both have two very different views on how to use and legislate artificial intelligence. Elon Musk wants to regulate the use of AI, in his words:

(AI) presents fundamental risk to the existence of human civilization.

According to Elon, you won’t have to spend much time before we see robots killing people down the street. (If you are thinking on a Terminator Rise of the Machines scenarios, you are wrong, I’ll write about this later in a new post)

Marc Zuckerberg has another completely different view, MZ says:

I’m optimistic. And I think people who oppose and try to paint these apocalyptic scenarios, I just don’t get it. It’s really negative and, in fact, somehow I think it’s pretty irresponsible.

Beyond all this, what is clear is that all the biggest Tech companies in the world are investing (one way or another) in something related to Artificial Intelligence. Every little advance that is made on the subject will affect the way we live and work.

I am almost sure that my children will not have the need to learn how to drive. In 10 years, cars will be electric and most will be autonomous. Depending on where you live, you probably don’t have the need to learn how to drive, because a car (which doesn’t have to be yours) will be the one to help you move when you need to travel from one place to another.

Scenarios like the previous one, are not far from being a reality. And what today seems to us science fiction, in a few years will be our day to day.

Greetings @ Burlington

El Bruno