
In this post (the first of a series?) I want to share some thoughts, insights, crises, and everything in between regarding the first 9/10 months of my new PhD, hopefully something useful for fellow students who want to undertake this path :).
I will begin with a short bio about myself, then I will go through my PhD story so far, and will conclude with a set of hints that I would’ve known at the beginning of this journey.
Note: I won’t use any fancy AI tools while writing this flow of thoughts, so please forgive any spelling/grammar errors, but as you will read at the bottom of the page, it’s a new style and life choice.
The Beginning
Let’s start, indeed, from the beginning. I moved to the Netherlands in late September 2024, landing at the Eindhoven Airport in the afternoon.

In the first month, and I will be quick in this, I went through the whole process of moving to the Netherlands, i.e., buying a bike, registering with the city hall in Tilburg – where I live – finding a doctor, registering with the health insurance, and bla bla bla. Let me know if you want a detailed guide on this process, especially with finding a place to rent, which is by far the toughest part.
But what am I doing here? I graduated in April 2024 from the Polytechnic University of Milan with a degree in Computer Science and Engineering. After a 6-month internship in an R&D department in France (cĂŽte d’Azur, wonderful place), I figured out that I really like research and having an impact on society. For this reason, I searched for a few positions in R&D in big companies, where my research could’ve had a big impact. I found out that most of them have a PhD degree as a minimum requirement. What I did, then, and I highly suggest this to anyone searching for a PhD, is to contact the professors with whom I had more contact in my master’s, and I asked for any available positions with them, or with anyone in their network. I was flooded with emails and positions. Big finding: there are more PhD positions than students willing to do a PhD. A small incise on this: I’m talking about Europe, and I am not aware of the situation in the rest of the world. So, do not stress too much about the interview or the selection process; they will very likely hire you anyway. But be very proactive, they love this! Moral of the story, I found a position at JADS (Jheronimus Academy of Data Science), a joint institute between Tilburg University and the Eindhoven University of Technology, having basically a position in both of them, but working in the beautiful Marienburg Campus in Den Bosch.



My PhD So Far: a Cold Start
It took me a while to understand what my PhD is about, and I am not fully convinced yet. But let’s start from what I knew since the beginning: I work in the Safeguard Lab, and my goal is to achieve, to some extent, trustworthy AI from an accountability point of view.
In the first 2 weeks, I wrote a vision paper (basically 6 pages of brainstorming), gathering some resources from the literature and experts in the field through a quick and dirty survey. The paper got accepted in the PhD Symposium at ICSOC 2024, and it’s available here. I met very interesting people at the conference, and this is something you always want to do; it’s the whole point of a conference, in the end. If you are kind of introverted, tho, like me, you should think about going with fellow PhDs from your group or with your supervisors. Unfortunately, this time I was alone, but I met a group of Italian professors and PhDs with whom I spent almost the whole conference. It’s nice because you can really socialize a lot and do some networking while exploring the city or doing fun activities (besides attending the conference, of course). For example, we’ve been to the Medina in Tunis and to a very fancy hotel for a cocktail and a lot of very nice Tunisian food.

On another day, we went to visit Carthage (old Roman and we had another nice dinner.

I highly suggest attending a PhD symposium to anyone starting a PhD, as this is really a great start for a bunch of reasons:
- You get to do something as soon as possible.
- You get to know the people at the conference (you will likely see them for the rest of your PhD at many conferences).
- You define a set of preliminary research questions. This will guide you through the entire first year; the topics will be very broad, but at least you have something to work on from the beginning.
- If you get accepted, you experience something new and super funny after 2/3 months from the beginning of your PhD. This is great, as this is usually the moment where you start asking yourself: “What am I doing here?”.
This question followed me from the 3rd to the 6th month. This is where I tried everything: literature reviews, interview-based study, proof of concepts, etc. More specifically, I tried to search in the literature for something interesting, as this should be the first thing to think about. I didn’t have a great success in this; it’s really hard to find something if you don’t know what you’re searching for. So I asked some of my colleagues, supervisors, and professors I knew around me, for something to do, and if they needed help in finishing a paper or project (be careful if you want to follow my steps: they have a lot of stuff to do, don’t say yes to everything because then it could be too much). This kept me busy for a few months, and in the meantime, I could keep exploring my research field and search for something that met my taste.
Many people in my lab are experts in Data Meshes, so I tried to dip my toes into the subject by joining an ongoing work on this subject (you can read more about the topic here https://www.datamesh-architecture.com/). The work is an interview-based study conducted with several Data Engineers in companies all around the world. With this work, I understood if I liked the topic (I do) and if I like this kind of primary and qualitative studies (I do). Although I’m not very sure if this research (or the following ones) will contribute to my thesis, I believe it has been really important both from a team perspective as I got to work closely with other people in my lab, and it’s something that will end up in my CV, at some point, when someone will accept the paper đŹ (this last point will need more space, and I will write a post about my perplexities).
Among these collaborations, I got to work with a professor from my previous university in Milan (Politecnico di Milano) and my supervisor in the Netherlands, on a project about the integration of Zero-Knowledge Proofs with Machine Learning. While writing a paper on this project during one of the many sprints for a very important conference (ICSE 2026), I figured out a couple of interesting things:
- The project has been stalled for too long, and a lot of people have already published on this.
- I could rewrite it to match my PhD thesis and research question!
This was really a Heureka moment: I found something I liked and I could write a couple of papers to be included in my thesis!
In this exact moment, is where the Dutch bureaucracy came in handy: my university wanted me to write a Research Proposal, a very long and detailed document that describes my project, the problems I aim to explore, possible solutions, and deliverables for the 4-year planning of my PhD. This turned out to be very useful as it forced me to think about my planning and to finally find an answer to the question: “What am I doing here?“.
After this, I produced what I think will be the second and third chapters of my thesis. The first one is a survey-like study (submitted to Transactions on Software Engineering; the non-peer-reviewed preprint is available here), and the other is an actual framework that aims to include ZKP auditing into MLOps practices (submitted to ICSE 2026).
In June 2025 I attended my second conference with two posters that I co-authored with other colleagues. The name is SummerSOC, the summer version of the ICSOC conference I attended in Tunis. My supervisors were both closely involved with the organization of the conference, so they really wanted us to take part in it. It has been a really nice experience, and it’s been a good opportunity for some nice team bonding in a very nice place.

As you might have understood, conferences are my favorite part of the PhD. You see new places, meet new people, and hang out with colleagues and friends. Submit as much as you can! But don’t forget about Journals, they’re more important!!
Next Steps
I already have a couple of interesting things to do in my next 6 months. I need to finalize my papers following the reviews (and possibly get them accepted), and I will join a new interesting project: Marit-D!
In this project, we are collaborating with the Dutch police, as well as with the French, Portuguese, and Spanish police. The goal is to catch anomalies in the sea. This project will serve as a case study for my proof of concept, too! It’s a high-risk application, and compliance and safety are crucial. If we get to apply my framework to this scenario, the third or fourth chapter of my thesis will be done!
I am gonna post more details about my ongoing projects in this series, for me and for those of you who might be interested in following my journey.
Some Advice
I want to share a couple of advice that I wish I knew before starting my PhD. They are mainly related to going to the office on a daily basis and the use of AI for scientific endeavors.
I believe going to the office is crucial. Science is made not only by conducting experiments and writing down the results, but also by interacting with peers, exchanging ideas, discussing problems, etc. I wouldn’t have published anything so far if I didn’t have my colleague Renato introducing me to topics that I couldn’t have known without talking with him. Participation is what makes a university a great incubator for ideas and progress. I highly suggest that any PhD students go to the office at least 3/4 days a week. You will see the results.
Now it’s time to address the big elephant in the room. My PhD is in AI too, and I’m submerged every day in papers and news about it. As many people did, I got carried away at the beginning of my PhD with the (ab)use of AI for brainstorming, writing, fixing errors, everything. At the end of the day, tho, I felt like going backwards and not learning anything. The moment I decided I would stop using any AI (or LLM, to be precise), that’s where my PhD started taking shape. I had more original ideas, more effective communication, and I felt like I was actually learning something. It was just slowing me down, and I felt like I had no control over my projects. The same applies to coding; I completely stopped using it, and my software became just better. Here are a couple of sources to prove my point that I highly suggest reading:
- Why LLMs Can’t Build Software – this resource explores the concept of mental model into software architectures.
- Early 2025 AI Experienced Open Source Developer Study – this study shows that when developers are allowed to use AI tools, they take 19% longer to complete issuesâa significant slowdown that goes against developer beliefs and expert forecasts. This gap between perception and reality is striking: developers expected AI to speed them up by 24%, and even after experiencing the slowdown, they still believed AI had sped them up by 20%.
I’m open to any discussion about this last point, so feel free to leave a comment below.
Acknowledgements
I would like to spend a paragraph to thank the people without whom I wouldn’t be here writing this post. First and foremost my supervisors for allowing me to begin this great journey. A huge thanks goes to my family and friends for supporting me in every small step that led me here.
Last, but not least, I would like to thank ThĂ©mis. Thank you for making me feel a bit at home, even if sometimes can be difficult. Your support is a fuel for my work, life and everything around it. I hope to be your fuel too. (I made a promise to you: I won’t let the AI take your beautiful job, the kids need a wonderful teacher like you).
Thank you for reading the post! Feel free to leave a comment down below and to reach me out /contacts.

Comment
U want to go out for a date? xx