My perspective after two years as a research and teaching assistant at FIDIT
My employment as a research and teaching assistant at Faculty of Informatics and Digital Technologies (FIDIT for short), University of Rijeka (UniRi) ended last month with the expiration of the time-limited contract I had. This moment has marked almost two full years I spent in this institution and I think this is a good time to take a look back at everything that happened during that time. Inspired by the recent posts by the PI of my group, I decided to write my perspective on the time that I hope is just the beginning of my academic career.
Motivations and goals
I finished my master's studies in physics and informatics at Faculty of Physics with a master's thesis in quantum physics education titled Conceptual approach to selected contents of quantum mechanics. I was very interested in further inquiry into building virtual laboratories for quantum physics, which would serve as an accessible resource for primary- and high-school students while they are learning the fundamental concepts of quantum mechanics. However, I soon found out that it is not possible to choose physics education research on the Ph.D. level at the Faculty of Physics.
I started to think about my topic of interest as a great example of how we can use informatics to solve a problem in the education of (quantum) physics. I decided to inquire further at FIDIT and see if there is an option for enrolling in a Ph.D. study of Informatics with this or a related topic. After some discussions with potential mentors, I learned that virtual labs are already commercially available and it is not going to easy to find a Ph.D. topic in an area that is so mature (formally speaking, as I learned later, virtual labs have high technology readiness level).
You can hear all kinds of horror stories from Ph.D. students: how it is too hard and you have no time for life, how you end up soulless and drained, and how you slowly give up on your hopes and dreams. In my case, however, the hopes and dreams I had were crushed even before I started working on the Ph.D. thesis.
I went back to the drawing board, almost literally, and decided to look into the areas and topics offered by professors at FIDIT. There were several interesting options to choose from, and the one closest to physics was simulation of molecular dynamics. Despite being familiar with atomic and molecular physics, I felt that I wasn't ready to take on this topic in my Ph.D. Unfortunately, during my bachelor's and master's studies, the subjects of physics and informatics were seldom connected in these terms. To put it simply, I could imagine creating a virtual three-dimensional laboratory in Unreal or Blender, but using parallel processing to compute forces between interacting bodies (atoms or planets) seemed so far away from everything I knew at the time.
However, my (at that point, still to be) mentor reassured me that it was no problem, that my background was sufficient, that my motivation was of the right kind, and that it would all work out fine over time. Finally, in October 2021 I enrolled in Ph.D. studies of Informatics and in December 2021 I started working at FIDIT, where I joined Group for Applications and Services on Exascale Research Infrastructure (GASERI or 😎 for short).
Scientific research and networking
In addition to attending Ph.D. courses, I started my research work that plans to address some of the challenges in computational biochemistry in the exascale supercomputing era. In my first year of Ph.D. studies, I:
- took two courses organized by Partnership for Advanced Computing in Europe (PRACE) courses, HPC fundamentals for end-users and HPC - Intermediate Concepts for end-users,
- attended BioExcel School on Biomolecular Simulations with a poster presentation titled Dynamic computation of atom weight using expressions in GROMACS pull groups,
- attened Science and communication workshop organized by UniRi Doctoral School,
- presented my first paper titled Towards General-Purpose Long-Timescale Molecular Dynamics Simulation on Exascale Supercomputers with Data Processing Units at 2022 45th Jubilee International Convention on Information, Communication, and Electronic Technology (MIPRO) in Opatija, and
- co-authored the presentation titled Extending Non-Equilibrium Pulling Method in GROMACS with Arbitrary User-Defined Atom Weight Factor Expressions that was given by my mentor Dr. Vedran Miletić at Third Infinity 2022 in Göttingen, Germany.
I liked all of these experiences very much, especially those that were held on-site since they provided great networking opportunities. I mostly met fellow doctoral students (from UniRi and elsewhere), but also master students and postdocs. This provided me with an overview of what scientific problems are various researchers and their groups trying to tackle and how my research fits in the picture.
While it is still quite early, I started to think in terms of opportunities for collaboration with other researchers and groups in the future, especially after I finish my Ph.D. studies. I would like to keep my options as open as possible for the choice of postdoctoral research topics since it is hard to say right now what will be available to me in a few years. However, I am just looking as I don't plan to start any big collaborations outside my Ph.D. thesis topic for some time; I would like to stay focused and finish my Ph.D. studies first.
After passing my qualification exam, in the second year of my PhD studies, we also took part in Božićologija and Mutimir. I started keeping in touch with our collaborators and already have some interesting results from my research on molecular dynamics simulation methods, which is encouraging. I will publish them as a part of my first journal paper I am still preparing at the time I am writing this post.
Finally, I also participated in the MESOC project, where I helped with the preparation of the input dataset.
Teaching and administrative duties
I was initially assigned laboratory exercises in Computer Networks and auditory exercises in Network and Mobile Operating Systems. Later I also got auditory exercises in Computer Architecture and Organization. For all three courses, I did my best to reuse and improve the teaching materials I inherited, filling the gaps where I saw them. In particular, for Computer Architecture and Organization, I started converting the existing materials to Markdown, rewriting and expanding most of them in the process, to include them among the teaching materials for other courses.
I noticed that FIDIT is very much focused on teaching and it takes a lot of time and effort. Most of it is not even spent preparation of materials but on everyday support tasks: writing syllabi, preparing exams, grading homework, answering e-mails from students, etc. However, the preparation of at least some new materials is necessary nearly every year; the field of informatics is evolving rapidly, the software stacks are changing, and even the versions of the same software are changing the usage patterns, behavior, limitations, and feature set.
For an example of such changes, consider Docker, which is used for roughly half of the course in Network and Mobile Operating Systems. Docker Compose V2 replaced V1 and BuildKit replaced the legacy builder so the exercises had to be updated to follow these changes. This is far from the only course, as there were significant changes from the legacy to the current graphical user interface of the Common Open Research Emulator (CORE), which happens to be extensively used in Computer Networks laboratory exercises; they will have to be updated at some point as well. I am well aware these changes are not made deliberately to make life hard for teaching assistants at universities and I am sure that CORE developers don't come to work with great passion to make breaking changes to the way software is used. I noticed that they kept most of the links intact when they changed the documentation builder from Jekyll to Material for MkDocs, which is good for our teaching materials that link to them. By the way, interestingly enough, I recently did the same stack change (replaced Jekyll with Material for MkDocs) on my professional website.
Don't get me wrong, I like studying operating systems and computer networks very much and I have invested a significant amount of time into practicing system administration on real-world problems. In particular, I helped with maintaining the computer classrooms on several occasions just to learn how to work with Ansible, pacman, and systemd. I think it is great in general to stay up to date by learning new technologies and then be able to teach them, but I would prefer to have more time to focus on doing my research, where knowledge of these technologies is just a useful support skill. Understandably, FIDIT has to focus on teaching to help students benefit from the growth of IT jobs, but it is quite unfortunate for me that most of the paid positions for Ph.D. students of Informatics require partaking in teaching duties.
Time for tears?
So, since I enjoyed the experience, can I say that I am sad to be leaving FIDIT? After all, I would have stayed if I had been offered a contract extension, but it did not work out that way.
There will be no tears from me. It is what it is, and it remains to be seen what the future will bring. I anticipated this particular outcome and already applied for several positions during the last few weeks, most of which will allow me to finish my Ph.D. studies alongside work. Perhaps it will take a bit more time, but I am getting more and more interested in the topic now that the first results have started to appear. Also, I would anyhow like to finish the Ph.D. thesis after I have already invested two years in researching this field towards the thesis.
In the end, I would like to thank FIDIT for giving me this opportunity to explore my interests both in research and teaching and my mentor Dr. Vedran Miletić for the help and support he provided me thus far.