I sometimes forget how important passion is for engineers, teachers or scientists (anyone really)
“I sometimes forget how important passion is for engineers, teachers or scientists (anyone really)” LinkedIn Article
Denis Rothman’s book title: Transformers for Natural Language Processing
Hello, my name is Raul, and I always try to learn from my mistakes.
This may sound like a classic niche sentence, right? Anyone else may say, who does not?
Well, I learned to be a loser to learn from what I did wrong, which, honestly, has happened more times than I could count with my bare hands.
Then, why do you feel passion is as important?
Let me tell you something: it is a natural step for a human being to strive for perfect outcomes as they practise something more and more. Obviously, we get frustrated when things go wrong or when we are not successful in our purposes, whether in life, at work, or in our studies.
For example, last year, after a sudden round of lay-offs at my last company, I got the suspicion that I needed to change what I was doing so far. Things began to change swiftly because of the new AI bubble. However, I started this process of change a while ago, and for the first time ever, I started studies to perhaps become a teacher at some point. Nevertheless, this was not enough for me.
Therefore, a new bell rang in my brain; something told me to try again to join an MSc in Artificial Intelligence, for which I had reached second place in the queue the previous year. So did I. My partner told me that it was going to be so demanding, and I may get really stressed out by it. She could never have been more right.
Then, last September, I joined this MSc on Artificial Intelligence, Pattern Recognition and Digital Imaging internally at Universitat Politècnica de València (also as an MSc’s Assistant Class Representative). Could you imagine me? More than 10 years after completing my MSc at Cranfield, I am now studying a new Master’s in Engineering. So things got really harsh back in October, where my bday is… I had real anxiety because we had 4 modules with 4 presentations, 4 exams, 4 full assignments, and all to be completed within a couple of weeks after only 2 months. I am not gonna lie to any of you, I was about to run away and leave it. I got really frustrated because there was no time to study, and all tasks had to be completed in a hurry, and assignments accounted for at least 80% of the final mark. So I had to play my cards and leave the exams alone, and I got a 0 for the first time in my life. However, one of the teachers was great at supporting us and helping us become successful. Regardless, he held weekly group sessions on Pattern recognition and the machine learning module. Thanks for all your insightful support at “tutorias”, Alfons Juan Ciscar.
No matter what you may feel, if I tell you that the final marks were not bad, it was the initial frustration that was the problem. With my partner’s help, she convinced me not to leave it and to try to fight and push harder. Also, a special mention to my colleague on one of the first assignments for the Computational Linguistics module, Germán Rodríguez Díez, who is an outstanding scientist and a better person.
In addition to this complex situation, if I have not mentioned it so far, I was still studying my Master’s in Teaching and starting my internship at GRUPO SOROLLA EDUCACIÓN back in September 2025 to complete my compulsory hours in a secondary and vocational training course. Honestly, this experience was grand, and I enjoyed it so much both because of the experience I had as a teacher in practice with the “FP Superior en Administración de Sistemas Informáticos en Red” and “FP Medio en Sistemas Microinformáticos y Redes” students, and, obviously, because of my great colleagues at work, special thanks to Lidia Contreras-Ochando, Daniel Stanislowsky (my tutor), Daniel Caballero López, Patricia Rego and Francisco Rodríguez Rodilla.
After all, those months passed, I completed my Master’s Thesis in Teaching using the Flipped Classroom methodology in my favourite programming language, Kotlin… my last exam for this master’s was back in December for “Procesos y Procesos Educativos”… the final thesis presentation was back in February 2026, and then one less master’s to count. However, the MSc in applied AI, NLP and pattern recognition continued…
Surely, I did well in the Human Language Technologies module, my colleague Jaime Perez Navarro and I created for the first time our own transformer for natural language processing (NLP), special mention to Professor Antonio Miguel Artiaga for the lessons given. There I also achieved a mark which certainly was much better than expected on both exam and assignments… my passion to always evolve and learn helped me overcome some of my fears, that imposter syndrome that I brought with me due to my lack of studies for a long while, my forgotten mathematics and… some of the most brilliant colleagues in class, who were best in class, which in my opinion I would never be able to compare with.
On the other hand, things continued their own flow, and more group assignments started for other modules, one of them about “Early Sepsis Detection and trade-off in models” with Óscar Alcarria Laserna and Pablo García López as my partners in crime for the Pattern recognition and machine learning applications module. A project I pretty much enjoyed, where I learned from brilliant folks, who challenged me with really uncertain and tricky medical topics to master. There, we introduced an innovative explainability layer on top of our XGBoost model, which helped us to predict sepsis from a medical dataset from the PhysioNet/Computing in Cardiology 2019 Challenge. To be honest, it went pretty well, we used SHAP for this explainability. Special mention to the teachers Jorge Civera Saiz (also for Machine Translation first LLMs assignments with him and Albert Sanchis Navarro), Adrià Gimenez Pastor, and Carlos Martínez Hinarejos.
Moreover, the Applications of Computational Linguistics module included interesting labs on detecting sexism and hate speech in tweets, memes, and TikTok videos. Both my colleague Giuseppe Di Palma and I learned pretty interesting things. Not only that, but we also competed in the EXIST labs at the CLEF conference. Again, passion drove us to learn how to handle failures in our first models and bad classifications in the validation data split. Long story short, our cascade model learned the best combinations coming from annotators and their physiological signals (Eye Tracking, Heart Rate and EEG signals), and combined it with a pre-trained encoder (BERT) model and fused some frames extracted with whisper large to extract based on the timestamps, which frames were meaningful for the model to predict not only if sexism or not was there, but also what intention the videos had, what categorisation they were into. After all, we achieved first place in the soft-soft evaluation (LeWiDi paradigm), which allowed us to present a paper after the module was completed, outside the scope of the Master’s. Thanks to that, today I am happy to say that the loser who had no idea about machine learning back in September will go to the CLEF conference in Jena, Germany, to present the paper, if the committee are happy with it, presumably as a talk and maybe a poster. Special thanks to David Gimeno-Gómez and Paolo Rosso for their support throughout labs, reviews, presentations, and for encouraging us to improve all the time. Also, to Roberto Paredes and Alberto Albiol for their insightful lessons on Artificial Neural Networks (here also to David Gimeno-Gómez) and Computer Vision, which helped lay stronger foundations for the computer vision bit I really needed for this project.
Also, one of my passions led me to complete two of the individual assignments, and I am more proud of them than anything else. On the one hand, in Metaheuristic techniques, I completed this running planner for half a marathon: running-plan-generator
On the other hand, in Data Visualisation, I completed this nice website with statistics from my lifetime of running: running-web-streamlit. I called it “hogar del corredor”, and this is the production website. Special thanks to the teacher Javier Lluch Crespo. Another interesting project was the first ever video game I created on the Computer Graphics module. Let me introduce you to “Animal Maze” (thanks to the practices teacher Mariano Banquiero).
In the final remarks on hiring processes, I humbly learned how complicated it is to prepare for interviews; you need to study, tackle challenges, and address complexities you may have forgotten from previous roles. I have to be honest: I failed at more than 30 applications simply because I tried to apply for roles where, normally, either you want a junior engineer, or my profile was too junior or too senior at the same time. How many rejections did I receive last month? More than you could imagine. On the Android roles, I interviewed with great companies and had poor experiences; some assignments I passed, others where bias could have been part of the process. In engineering roles in applied AI, there are engaging conversations. Again, long story short, it did give me very interesting lessons to learn, because you never know how much you need to prepare until you are under pressure for an interview schedule.
I do have to point out another remarkable thing: I had my first real teaching experience as part of the “FP de especialización” in the Cloud as well. I was lucky enough to work for a short period of time for GVA on the IES Juan de Garay with great colleagues such as Joaquín Casillas Meléndez, Jorge Lopez Querol, Guillermo Domingo Martínez, Rafa Chavarría Bonilla or Mila (sorry if I missed anyone). Once again, passion helped me to encourage myself on this adventure back in April 2026.
All in all, I can’t complain about this whole year’s experience. It has been challenging, and now I may take a summer break to recharge.
I need to always write in stone: passion is the fuel that powers and moves everything on and on.