Here I start to sing
to the beat of the wind,
that the man who unveils it
a sorrow extraordinaire,
like the solitary bird
with singing, he consoles himself
 

The text you have just read is from Martín Fierro, a narrative poem from 1872; when I read it, it was clear that my future would be to research the figure of the gauchos and teach Spanish-American literature. I didn't come even a little bit close.

I studied Hispanic Philology at the University of Seville. It took me longer than I had planned: I left the degree a couple of times for personal reasons, and when I took it up again, I realized that I had to go for something that would guarantee me a job when I finished (and as soon as possible). Generally, when you enter Philology, the career path you have in mind is to become a language and literature teacher or, failing that, a teacher of Spanish as a foreign language.

I was lucky enough to meet my great teacher and current project partner: Juan Pablo Mora. His classes, which seemed the most chaotic in the world to me, gave us the freedom to create projects on our own. He was the one who told me about Computational Linguistics and since then, in my free time, I read about it. Juan Pablo made me look at linguistics with different eyes and became my tutor for my Final Degree Project on linguistic manipulation strategies. When I finished my dissertation, I thought about the idea of automating these processes and being able to detect these manipulation strategies through social networks. It turned out that it already existed and had been done for a long time. It was then that I decided to enroll in the master's degree in Language Sciences at the UNED and apply it to the computational linguistics track. 

The summer after I finished the master's degree, I received a message from David Munárriz, Emergya's Director of Strategy and Solutions. He wanted to interview me to join the company. I remember something he said to me that has stayed with me until now: "I don't know how, but I know you can help us". Every time I face something new lately, I feel like I'm getting too small, but then I hold on to that: "OK, at the moment I'm not sure how or what, but I'm sure I'll find a way to bring what I know to this new challenge". 

Currently, I work in the linguistics area at Emergya developing virtual assistants and projects based on Natural Language Processing (NLP). 

As far as assistants are concerned, we linguists work on the cognitive development of the assistant on a day-to-day basis; does that sound complex? Perhaps this will make it a little easier to understand:

• We implement, according to the assistant's personality, the variational features (according to their social position, age, etc.); then, we transfer these responses to Dialogflow taking into account the variables that may occur within the conversation.
• We help define the conversational flows: intentions, dependencies... and validate it with the architecture department.
• We create training sets so that our assistant can understand us. These training sets are a vital part of our work since we have to define well the syntactic and semantic logic that we will use in each intention we want to capture.

This is just an example of what we do, but the truth is that in our day to day we are involved in other processes that require our work, such as data extraction, analysis and creation of NLP training models that we carry out through corpus linguistics. 

And it is not isolated work: our teams are multidisciplinary, with professionals in conversational design, development and architecture. Together, we look for solutions to the challenges we encounter, sharing and understanding each point of view. That's why communication and synergy between colleagues is vital.

Apart from this, I am currently involved in a research project for the US. This project is about the biases that are transmitted through Natural Language Processing. By generalizing certain patterns, we see how our speech biases are transferred automatically to AI. It is an extremely interesting project, and I feel very fortunate to be learning alongside great professionals.

Also in December, we started an NLP cooperative learning course for teachers and student interns at the University of Seville. This project is, once again, a collaboration with the US Department of Linguistics and @Lingunaria. The latter is a personal initiative that has just started this month. 

In January 2021 and with the help of Emergya, the US Department of Linguistics and other colleagues from the Biases and AI project, we started the project: "Young people with researchers", a monthly conference to bring our work closer to students from high schools in the province. In this case, the project we will share will also be about biases in NLP and their impact on AI. We have already outlined the activities calendar. There are some attractive activities, for instance, learning how to use some tools for corpus processing. There are also others, such as identifying segregative patterns in current speeches on Twitter or Twitch. 

As you will see, being a linguist nowadays is a very demanding job. It goes from collaborating in transversal teams in virtual assistant development to analyzing and predicting behavior in social media and to cognitively 'improve' Artificial Intelligence in areas as complex as the elaboration of our messages and their impact. It is a career that, thanks to technological advances (something that may seem paradoxical at first glance), is very fashionable because, in a world where automation marks the day-to-day, humanistic work is more necessary than ever.