I just came out from an event of Entrepreneur First: the final session of the AI StartUp School of 2025. My decision of quitting my PhD was taken a few months ago, but a very interesting discussion I had with Matéo Petel during this event made me realize that I have to write about my decision somewhere. As he said, writing is an act, and a beautiful one, and we are lucky to be able to read someone's mind. His words resonated deeply in me and changed my perspective on how I was seeing writing and reading. I would say that I was pretty much biased by research and had forgotten the essence of writing and why humans do it. So thank you, Mateo.
This article was also a little bit edited at the airport of Agadir at the end of a surf trip in Morocco while finishing the last lines. In this article, you will see that traveling was very important in my decision process, and now I feel more confident than ever in my decision to stop my PhD and truly embrace life.
I think I should start this article from when I signed my PhD contract. I was freshly graduated from the famous master MVA in Paris, the greatest path to AI in France. You can meet there a lot of researchers from DeepMind, Meta and so on. I had been doing my internship at Kayrros, a very nice company doing intelligence with remote sensing technology. They were working very closely with different labs. I worked with the biomass team on generating land classification maps from Sentinel-2 and Landsat-8 sensors by training U-net models. I really liked the internship; people were smart, with a very good balance between research and production. The latter being really important for me because it constrains your research and obliges you to use the most adequate techniques, and you can see your work being used to solve real problems. It feels like accomplishment.
At the end of my internship I had the choice to either continue with a full-time contract with them or to continue in a PhD on similar topics. By that time, I wanted to go deeper into the technology and do a long-term project. I love that; I am very keen on taking complex challenges that put me in very long-term projects, with a lot of training involved to get better and learn. For instance, I attained very good levels at judo, athletics, and e-sports while younger. Grinding is something I really like. And so the PhD seemed to be a very good place to do that.
The subject we created together with my two supervisors (well-known researchers, I also thought this was a good hint of a good PhD, but let's discuss that later) was very appealing: detection, segmentation and classification of forest disturbances from satellite imagery. Honestly, the setup seemed very right at that time. I would be between two great labs: LSCE and INRIA, and also would be close to the Kayrros team which I was used to working with and which was interested in the topic.
The PhD really started well. In 6 months I had made great advances, met with a lot of researchers from France, Germany, Denmark, and even presented my studies in front of NASA researchers. This was great and I was learning a lot. After the first 6 months, I started to be annoyed/blocked by the lack of data and how things were working between researchers: it was very slow to get my hands on the data, and being between two teams was not very great. People at INRIA were doing maths, people at LSCE were doing climate science, so I did not fit in either of the two.
People from LSCE were kind for sure but kind of lone wolves working in a lab very far away from Paris, cut off from "normal people". The lab was 1h30 away from Paris, so that means spending 3h commuting to go to work. I did not go to the lab often... In the other lab, INRIA in Paris, people were also cool but doing pure maths. And globally, I was not feeling at the right place when surrounded by researchers; people were talking all the time about the same things and lacked vision and life. It feels like something was turned off in them. You know this spark of life that some people have? I never encountered any researcher having it, I felt very sad about that. I was afraid of becoming like them. I know it is horrible to think like that, but really I was not inspired by them. And each time I was going to the lab, I could see this and I was feeling very different. I was very well welcomed in both labs but did not feel at ease there.
So I decided to work at Kayrros from time to time to be close to people having great experience in both research and engineering. In this topic, I wonder if having great knowledge in engineering isn't the key: you handle a lot of big data and various types, so you have to be a bit of a data engineer, especially if your team doesn't have any pipeline ready for you. Six months into my PhD, I was doing more engineering than research because I was exploring things that my two teams had not covered yet. Thanks to my experience at Kayrros (and ChatGPT), I was not that bad at coding and could find workarounds to almost all issues. It felt great, honestly, completing things that both teams couldn't do. It felt like a challenge, but this was slow, and I wasn't really doing research but preparing everything for research... asking for data, negotiating for data, preparing data, building pipelines for data, etc.
I was still confident at that time and thought that this was normal; this is a PhD, everyone says it's hard. Every Friday, I was sending an information letter to my supervisors, and they seemed happy with my work, so I had no worries. So I continued for the next 6 months, and struggles emerged increasingly. By that time, I was starting to understand that it would be difficult to end my PhD on time because I was chasing data and not doing Deep Learning at all. This was a big disappointment for me. I really wanted to explore deep learning, but I couldn't because of the data lacking. But again, I was confident; okay, it's the first year. After that, I would be able to start training models.
Without realizing it, I was spending a lot of time on my PhD and forgetting a bit everything around me: girlfriend, family, friends, and myself. I completely lost myself. After one year in the PhD, I broke up with my girlfriend: the PhD is not responsible for that, we weren't on the same page, but it is an important piece in the puzzle. So without any accommodation in Paris, I came back to my hometown at my parents' house to find something else in Paris. I thought this would be for one month, but the housing market in Paris was very harsh and I couldn't find anything soon. So I asked my supervisors to stay a bit longer in the south. Honestly, I enjoyed it. I was missing the sun, my family, and everything. Paris can be very harsh sometimes.
Far away from Paris and with less and less contact with my team, I was even more annoyed by my PhD. The topic was really interesting, but I was collaborating with no one. My first article was stuck because of data issues, and I was trying to find a workaround by creating a new method, but that was too complicated and would have been another PhD topic entirely. Taking a step back now, I made some bad decisions by continuing exploration at that time; I lost precious time. That was my fault, but I am a bit pissed at my supervisors for not having supervised me more since the beginning: not preparing the subject enough, following my track superficially. I felt that the PhD was not prepared at all and that I was doing exploration all the time. I wanted to do a PhD in Deep Learning and I ended up looking for data and preparing it. Honestly, I'm also pissed at myself for being so naive from the beginning and thinking that being with well-known researchers was a good enough sign that things would go well. I tried to be meticulous from the beginning by sending those weekly letters, and I thought my supervisors were reading them seriously. They weren't saying anything, so I thought (a bit naively) that everything was all right, that I was doing good, but this was not the case at all.
By that time, I was very disappointed with my PhD environment. I had the feeling I was doing work that no one cared about, even my supervisors, and just running after articles to have this fabulous h/i-index going up. I was stuck; I had the feeling I was losing a lot of energy in this PhD, for nothing. It was frustrating. I had plenty of energy, but I was not seeing the impact of it: no encouragement from my supervisors, no team spirit, it was harsh. So I decided to tackle another challenge to change my mind and to ease myself (and earn some money). I started to do some freelancing in my free time and I really enjoyed it. I needed a break from my PhD. I asked to take a year off to explore that and do a program like Entrepreneur First. They refused... On one hand, I understand that there was a possibility that I wouldn't come back after my break, so they directly said no because they thought I wouldn't return: they told me either stop now or finish your PhD. Obviously, I decided to continue.
I tried to be remotivated. I took a pause from my first article and went for a new, easier one. I also made great progress and had almost completed the second article. I just needed data again. I had data from an institute (Theia) in France, but the quality of it was just off. So again, I was looking for data. This time Sentinel-2 monthly mosaics over France. Data was scarce; the lab had enrolled more PhD students and one engineer who were working on a pipeline to download Sentinel-2 data (yes, 2 years after I started my PhD). This pipeline was not working well... I was stuck again. This time I was very frustrated; this was the second time I was planning an article and got stuck with data. So I developed a method to compute the phenology (deciduous or evergreen) of trees in France from Sentinel-2 time series. It was working pretty well. Then my supervisor asked me to add the entire map of France to have a good article. This was good, and I was going for it. But again, data issues. So I lost 3 months waiting for the lab pipeline to be ready to get the data.
Nevertheless, being stuck, I decided to join my cousin for one month in Bali. I needed it; I was really lost and frustrated. Always stuck by the data. Honestly, this break was so important for me: I reconnected with myself, with what I love to do in life, and I understood that work is not everything, indeed quite the opposite. Life is short and you have to do things that you love. Obviously, I was not liking my PhD so far.
At my comeback from Bali, I was very happy. I had lived a bit. It was two years into my PhD, and I decided to come back to Paris to be closer to the team and to really finish everything. I was motivated. I was saying to myself that this was the last run, and after that, goodbye to research. The last effort. Once I arrived in Paris, I was disappointed again: I wasn't seeing my supervisors more, there was no team... it was even worse. But the disappointment was my fault really. I'm someone super optimistic, and people from the team had warned me about going back to Paris—that the PhD environment would not change much. People at the lab were conscious about that and suffering too; I was not the only one who felt lonely or who was expecting other things from a team.
At this time, I was lucid enough to say that I didn't want to live like that anymore. I loved the topic but didn't want to be in such an environment. I was not happy and not progressing in life. I lived in Bali for just one month and had seen plenty of happy people. I was just depressed seeing people in the labs, lifeless. When I made my decision, I was very harsh on research and researchers—I was angry at them. But I also heard very good stories from other labs with good teams, so I guess I just fell into a bad environment for me. Then the emotions faded a bit. Today, I think that my environment was not good, and my supervisors did not do their job at the very beginning to streamline the topic and warn me about things in this domain: people are super conservative, having data is complicated, and all this should have been thought out from the beginning. I really wish they had taught me how to do research properly at the beginning. I learned it, but the hard way, and it should not be like that.
Now I do think that I made several wrong choices starting the PhD. I am a hard worker and the PhD project seemed to be a good one to explore and learn. But I was not passionate about research before and was not very interested in it. I am curious but not passionate about this. I think I was lured by the excellence of doing a PhD without understanding that being a researcher is really a way of life. It's not for everyone; it is hard, and you have to be very resilient, very focused on your project. It can be overwhelming. I just want to enjoy life, and since I am a hard worker, I ended up in a pretty good school, doing a great master's and starting a PhD. But is it what I really wanted? I don't think so. During my studies, I always felt a bit off, motivated by something else. I am pretty sure a lot of people felt that too, and this is why I am writing all of this. I do think stopping a PhD or leaving academia is all right. It is better to do what you love, and it is a luxury that few people can afford. But if you've arrived at the pinnacle of studies without being really into it—not passionate, not crazy about it, and not picturing yourself as a researcher—then I am pretty sure you are a very capable person and you should do something that makes you feel good and fulfilled.
Deciding to quit was profoundly difficult. It meant confronting feelings of failure, courage, and responsibility. Ultimately, however, continuing without passion, meaningful support, or a clear purpose was unsustainable. I now plan to complete my current article, thoroughly document my work, and leave a solid foundation for future researchers.
Last word for research/academia: I love research and I think some researchers are incredible people. This is essential to society. But this is not for me. I know that plenty of people love doing it, have a very good environment with great teams and nice challenges to tackle. My environment was very particular.
For those who still want to pursue a PhD because you feel it, you are passionate, and you want to pursue academia: my advice for you is to choose a very good environment with a great team, and not to go for famous researchers. Those people generally don't have the time to teach you things, so go for someone who is ready to make you shine. I would recommend favoring the lab/team/environment/supervisor over the subject. Things are always cool if you are with the right people.
Today, I embrace a new path, grateful for the lessons learned and excited for what lies ahead.