Looking for a job can be a gruelling process, but also a learning experience. After some interviews (and many months of therapy), I feel ready to look back at my unorthodox academic journey. Sometimes, I feel that having taken many turns was a mistake. I would like to challenge that feeling. Without further ado, be my guest, dear reader, and accompany me on this trip through memory lane.

When I was younger, I had a hard time coming up with what major to choose. There was a program in high school to visit research groups. I was assigned to visit a research group in genetics. They were welcoming and always asking questions and bringing up new ideas. While most of the discussions went above my head, I felt like I wanted to be part of such a group. Naturally, I asked one of the senior researchers about which major to choose. She helped me narrow it down to 3 options: medicine, biology, or chemical engineering.

At the same time, I had other interests. Academically, I was doing well in language, maths and sciences. In my spare time, I was using emulators and playing computer games. My dad was well aware of that and brought my attention to other engineering fields in particular software, electronics, and biomedical.

Having so many options felt overwhelming; I didn’t know what exactly to pick.

Trying to make a balanced assessment, I chose chemical engineering. It had mathematics, sciences, and offered the chance to eventually allow me to become part of such a wholesome group of researchers in genetics by studying biochemical processes.

I learned a lot during my bachelor’s. I was exposed to both theoretical work, wet-lab, and even entrepreneurship, finance and management. This period was instrumental in helping me understand what wet-lab work actually entailed. I liked to analyse data, coming up with ideas and seeing how theoretical principles applied to experiments. The detective work of finding why experimental results did not match was very enjoyable. However, experimental work required additional and highly non-trivial preparation. Things like specialised pieces of equipment, reactants, and security (this one with very good reason) delayed the brainstorming part that I enjoyed the most. After at least 6 experimental courses, spread through 4 years, I decided that experimental work was not for me.

There was another area in which this detective work could be done without having to ask for expensive reactants or requesting a timeslot with a technician: simulation. Experiments taking up hours to set up and run became a couple of clicks to get almost instantaneous results.

At that time, my main tool was process simulators. These are specialised pieces of software that simulate what happens inside industrial chemical equipment. As I didn’t know the inner workings of those pieces of software, finding errors and correcting them became more of an art than a science. This stood in contrast to experimental work, where I could always go back to the textbook and know what theory to apply.

I got my degree, but I could not find a stable job. After 2 years of different jobs (one being in a laboratory), I got funding to travel to Norway and do my master’s in process simulation. This experience was eye-opening in many respects. The group in Norway knew how to use process simulators, but also how to build them.

It was awesome to see how they could code specific solutions from scratch. Up north, I got to learn many things about computers, like the basics of parallel computing, automatic differentiation, numerical analysis and even databases. At the same time, I had the chance to take courses (in cybernetics and the maths department) that showed me how mathematics was not just a set of instructions for computations, but a formal language to describe anything. In particular, I understood the value of proofs in an engineering context.

While up north, my thesis involved writing a simulation of a plant from scratch and getting data to guide the automation of said plant. It was my first time writing a medium-sized and open-ended software project. At some point, I got to face difficulties in getting meaningful results from these simulations. The final results were not the cleanest, and they left me wondering what I could have done differently. Emotionally, I noticed how my enthusiasm and energy went up whenever I was writing the maths and the implementation details. This was a whole different feeling from my lack of enthusiasm for the engineering interpretation of the results.

I felt super behind compared to all the other researchers in Norway who seemed to know the ins and outs of the maths they were using. They all were quick to recognise and apply mathematical concepts, while I was still struggling with understanding the basics.

Having to return to Colombia gave me a lot of time to think. So I took getting my maths knowledge up to speed as a personal challenge. At that time, machine learning was becoming popular, so it made sense to apply what I was (re)learning to machine learning.

With time, I caught myself going deeper and deeper into maths. At some point, I came across how differential geometry was used to describe common problems in chemical engineering. In my naiveté, I thought that could be the reason why there was so much numerical instability in the code of my thesis.

The sirens’ song was so enticing that I jumped the ship of engineering to dive into the sea of abstraction. I was not aware of what I was getting into. It took me 2 years to teach myself all the missing courses in pure mathematics, and I managed to get into the MSc.

This is probably the period of my life where I have seen the most beauty ever. As one Youtuber, said “I am surprised I didn’t burn out my beauty receptors”. I learned about many areas of mathematics, each with its own challenges and a special kind of appeal. As expected, I ended up focusing on differential geometry. Their techniques had plenty of practical applications, like control and thermodynamics. However, I never got into that.

After those 4 years of learning maths, reality struck me. I loved the discussions and the passion that all the people showed. Teaching was something I deeply enjoyed, and I put a lot of effort in making engaging teaching material. So it seemed obvious to pursue the academic path, do a PhD and become a professor.

In practice, it was not that simple. For starters, back then, I was deeply in debt. I had also been warned about the lack of stability in an academic career by both of my then mentors and more experienced mathematicians.

The economic situation meant that I needed to find a place where doctorates were well paid and did not require hefty application fees or expensive exams. So positions in Latin America and most of the US and the English-speaking world were discarded. Most of central and northern Europe fit these criteria. Having had a beautiful experience in Norway, I put most of my energy into finding a position in the Nordics.

The need for stability was more difficult to address. My then mentors strongly advised me to work on a plan B while doing the PhD. This could be either by doing some applied research or using my spare time to improve my programming and data science skills. Knowing what former PhDs did after their degree became a proxy criterion to choose a position.

To look for offers, I was consistently checking professors’ websites as well as mailing lists. My goal back then was to continue doing fun mathematics, and I was willing to compromise in cross-disciplinary projects. I had the chance to talk with several tentative supervisors, who sadly did not have funding at the time. They told me it could take anywhere from 2 to 6 years post-PhD to find a permanent position. They recommended that I have a plan B, in case life happened and I needed stability. This made me think deeply about what I really wanted, since by then I had only lived on temporary contracts for at least the last 5 years.

Leaving no stone unturned, I ended up on a website about maths and control. This website had a section for openings, where I found a position at Aalborg University that used many things from my previous experiences. It was a PhD in computer science. This seemed like an interesting compromise. I would be able to keep on doing some maths and at the same time get “computer science” written in my CV so I could have an easier time applying for jobs afterwards.

Back then, I did not understand what theoretical CS was about. Unsurprisingly, I had a big culture shock to see how things were done in CS. For starters, I had no idea about the foundational problems. Probably the most shocking part was the difference between communities. In maths, there were lots of free talks, short courses, and many people wrote and were eager to share free resources. In CS, the community felt more closed and results-driven. There were not as many people sharing free resources (or I had a hard time finding them). CS communication was done through conferences. The publication pace was around 2 papers a year, with strict deadlines and quick review processes. It then made sense to me that people in CS would not have much time to write free resources.

In retrospect, I was probably judging CS too harshly, as I was also going through some personal problems, and the whole world was opening up again after COVID. Still, I really appreciated the chance to code again. I was not aware of how much I missed it and how fulfilling it was to write programs and see them work. Due to my formal background in maths, I was assigned as a teaching assistant for courses related to programming languages. This made a huge impact on me, since it allowed me to combine my love for languages with programming and mathematics. Since then, I began reading more and more about programming languages and making prototypes in my spare time.

I was getting more and more second thoughts about continuing in academia after the PhD. I felt the decision could go either way. While preparing my defence, I saw a talk about a topic that sounded eerily similar to one I was learning right after my MSc in Maths. I looked into it, and it opened the door to a new world of connections between CS and maths. A part of me thought of this topic as a possible way to continue indulging my curiosity in academia. So, in a desperate attempt to keep the passion alive in a withering relationship, I applied for funding to pursue that topic.

Preparing the proposal took me around 2 months. It involved at least 5 rewritings of the same 2 pages, as well as asking many colleagues and friends for feedback. Also, out of inertia, I applied to several academic positions that seemed somewhat interesting.

Months passed, and I successfully defended my thesis. I was a doctor. One by one, I got rejections for all and each of the postdoc and funding applications I had sent. I felt… An immense sense of relief. Deep down, I knew I did not want to continue my life as an academic.

Looking back at this journey, I don’t think any of the steps were a waste of time. I am glad I was able to experience so many different disciplines. I wonder how many mathematicians can brag about the pain of having been through experimental courses. I feel lucky that in the buffet of life I was able to taste so many different dishes. One of the nicest feelings is that of getting a new mode of thinking (kinda like how Megaman gets a new weapon after defeating a boss) from each discipline.

Now I am applying only for industry jobs. It turns out that academia is plan B. The learning and the curiosity have not stopped. I have used these months to play with code and learn new technologies. After each interview, I do a debrief about the interesting ideas from the position, and those feed my backlog of personal projects. I have also learned a lot about how to communicate my skills in a more marketable way. Seeing the landscape of things to learn outside academia, I am sure leaving it is the right choice. The most important insight I got is that a nonlinear path is not a mistake; it is life.