 |
| My offices during the sabbatical. I had an office to myself at HPI (left); I shared an office with Matthias’ PhD students Arnab and Sebastian, who used to be my MSc thesis student, at TU Berlin (middle – photo credit goes to Sebastian here); at CWI, I was squatting in Hannes’ office (right). Sorry for the low image quality on the blogger platform. |
Previous semester (August 2023 –
December 2023) I was on sabbatical. This means that I had no teaching and
(almost no) administrative duties at the university.
This is a practical and personal,
not a technical, account of my sabbatical written for the general audience. If
you are interested in what I did scientifically during my sabbatical, there is
a bit of it here, but not much. I am happy to talk about that separately, and I
hope at least a subset of those works will get published eventually.
At ITU, in theory, you can take a
one-semester sabbatical after 6 consecutive semesters of teaching or one-year
sabbatical after 6 consecutive years of teaching.
Late 2020 (almost three years
after I joined ITU), I acquired my first research grant from Independent Research Fund Denmark. It was a
combination of a starting grant, Sapere Aude, and a grant targeting
junior female faculty, Inge Lehmann.
In my Inge Lehmann application, I put some budget for doing research visits
abroad since one of the things the call emphasized was the personal development
of the applicant.
As an academic, the idea of “doing
a sabbatical eventually” was of course in my mind, but I thought it would
happen sometime in the not-so-near future. At the time I acquired the grant, I
was still relatively new at ITU, this would be the first time I would get to
hire my own PhD students. Also, I still had a temporary residence permit and
didn’t want to complicate my residence rights in Denmark by leaving the country
for a longer duration with a non-EU passport.
Thus,
I wasn’t the one who asked for my sabbatical, it was my department head Peter Sestoft. During our yearly
one-on-one in 2021, he suggested that a one-semester sabbatical would be a good
use of my travel money from Inge Lehmann grant. His suggestion made sense to me
as well. So, it was decided.
Due to the teaching needs at the
university and my own need to first form my own group with the acquired grants,
the sabbatical was scheduled for Fall 2023.
I have seen people doing
different kinds of sabbaticals. The more traditional kind is when the person
goes to either another academic institution or a company to spark new ideas,
start new collaborations ... On the other hand, some stay where they are and
spend more time doing research with the absence of teaching and administrative
duties, some use the time to write a book or found a startup, some just rest,
some utilize it as an extended paternity leave, and some do a combination of all
these things. All are legitimate options in my opinion. The choice should be up
to what makes sense for you at that point in your career and life.
In addition to the obvious goals
like exchange ideas and start new collaborations, I aimed at the following
during my sabbatical: (1) minimize bureaucracy à
no new residence permits or visas – you have dealt with enough of those in the
last decade, (2) get to know different research groups better à visit different places
on the way, (3) introduce your team’s work to other people à give as many talks as
possible, and (4) remember that you don’t function well in not-crowded cities
and your mental health is more important à
prioritize work, but if you can, pick bigger cities for longer stays.
In turn, I decided to do a very
mobile sabbatical. I would first stay in Berlin for ~2.5 months splitting my
time between Data Engineering Systems
group at Hasso-Plattner Institute, hosted by Tilmann Rabl, and
DAMS lab at TU Berlin, hosted by Matthias Böhm. Then, I would have a 2-week
stay in Amsterdam at the Database
Architectures group at CWI, hosted by Peter Boncz. Finally, I would tour
Switzerland for 2 weeks giving talks at University of Fribourg (host: Alberto Lerner), EPFL
(host: my academic
mother), and ETH (host: Ana Klimovic).
While in Germany and The
Netherlands, I also took the time to give talks at other places either using my
own connections (Volker Markl
at TU Berlin, Zsolt István and Carsten
Binnig at TU Darmstadt, Martin
Hentschel at Snowflake-Berlin, Sebastian Schelter
at University of Amsterdam, and at TU Munich for the occasion of Lukas Vogel’s PhD defense) or with the
help of my hosts (Matthias invited me to the BIFOLD Summer School,
Tilmann helped me to get invited to the HPI Retreat, and Peter put me in
contact with Databricks in Amsterdam). The full talk itinerary and the talk itself can be found here.
While this plan satisfied all my
goals, nothing in life comes without trade-offs. The main drawback of doing
such a mobile sabbatical is that it becomes more challenging to focus on one
project where you can progress faster and dig deeper. Furthermore, visiting
several places in such a short time while other work duties go on in the
background can be exhausting both physically and mentally. Some days, I felt very
tired or couldn’t sleep well because too much was going on in my head. On such
days, I questioned my choice of splitting my limited time across many places. In
the end, I don’t regret the sabbatical plan I had, but I would like to be
honest about this trade-off.
Some people assume that
sabbatical means you are completely off your regular job and other work duties,
and for some this may indeed be the case, but not for everyone. At least, I had
to do other work simultaneously with all these visits.
First, I kept talking to my
students regularly online during my sabbatical. For me this is important,
especially since the PhD duration is only 3 years in Denmark, and I don’t
regret these meetings.
Second, even though I was off
teaching and administrative duties for the fall semester, I still had to deal
with the administrative stuff related to my course in the spring semester in
addition to paying attention to the open faculty call for our group at ITU. While this wasn’t a lot of
work and I didn’t mind doing it, it does cause some interruptions.
Finally, I had a lot of academic
service work. In retrospect, I probably shouldn’t have taken on that much. I
still struggle with finding the right balance with respect to academic service.
I enjoy doing this type of work and find it fulfilling and educational; the
amount tends to be the issue, not the work itself.
As a result, during my first
month in Berlin, I was working six days a week and could take only Sundays off,
and my email load wasn’t less than what it is during a regular semester.
Let’s talk about the individual
visits now; first, focusing on the work, then, on practical matters (e.g., insurance,
accommodation) and overall wellbeing.
Berlin
I worked with Tilmann and
Matthias before on different occasions. I knew that they would be open to
collaborating with me and there could be synergies across the research that we
do. Given the shorter-visits nature of my sabbatical, it made sense to reach
out to them first (late 2022) to kickstart new collaborations/projects more smoothly.
The way we approached the
collaboration was different at the two places.
Tilmann gave me time to talk to
his students the first couple of weeks of my stay in Berlin. Then, I told him
the project I find the most relevant given my research background and
interests, and we started collaborating on that one. I have more of a secondary
supervisor role in this collaboration. Marcel Weisgut, one of Tilmann’s
PhD students, drives the work. The project is on analyzing different
cache-coherent interconnects and allows me to get back to my transaction
processing roots, which I really enjoy.
With Matthias, I tried something
riskier. I haven’t been coding properly since I left IBM. I told him that I
would be open to getting back to coding. But it had to be in c or cpp (so no SystemDS for me). We discussed
different project visions he had, and I picked one. In this setting, Matthias
has been like my advisor, and he is a really nice one. (Not sure if I have been
a good student, though.) I am also involved in one of the BSc thesis projects
he has. Both topics here are out of my comfort zone as they are more into the inner
workings of machine learning. I work on machine learning systems now, but my
focus is more on the systems and hardware side. I have been learning the
machine learning internals very slowly and only at a necessary level in the
process. As for coding, I really enjoyed getting back to it, but given
everything else I was doing in parallel, I could only focus on it one day a
week when I was in Berlin, so the progress has been slow.
Tilmann, Matthias, and I also
discussed ideas for a joint project that we can kickstart in the near future.
This may or may not materialize, but I am optimistic and excited.
Amsterdam
In contrast to my history with my
Berlin hosts, I have never had a chance to work with anyone from CWI.
I have learned a lot from the work done by the Database Architectures group at
CWI, especially early in my PhD when I was trying to understand database systems.
They knew me from conferences. For the remaining part of my sabbatical, I
wanted to visit them to get to know the group and the place better and see if there could be
avenues for collaboration in the future. So, I reached out to Peter (during SIGMOD
in June 2023), and he was very welcoming.
I was aware that I wouldn’t have
so much time for this visit. After Berlin, I had roughly a month left in my
sabbatical before Christmas holidays. I knew I had to put Switzerland on my
path as well, but we will get to that later. Thus, I could only have about two
weeks for the CWI visit.
Two weeks is too short to start
anything from scratch. In addition, at that point in my sabbatical, my head was
quite tired due to splitting across different projects and locations. I unfortunately
didn’t have the headspace and energy to jump into something brand new while I
was still in Amsterdam.
Instead, I decided to use my time
to develop a few project ideas I had that involves using DuckDB; related to emerging SSD technology, resource
management on resource-constrained devices, and data management support for our
experiment tracking platform radT. I had very
helpful discussions with Peter, Hannes
and Mark at DuckDB labs, and Till from MotherDuck on these ideas. With the start of
2024, I managed to take the baby steps on a couple of these projects, we
will see how things evolve.
Switzerland
I did my PhD in Switzerland. People
have been inviting me to Switzerland since I moved back to Europe. While there
are a lot of people whom I consider to be family in Switzerland, I tend to
avoid visiting the place and prefer catching up with the people at conferences
elsewhere instead. My complicated relationship with Switzerland surfaces here
and there in this blog, most recently in this
post. I won’t delve into it further here. During my sabbatical, it made
sense to stop the avoidance and carve time to visit Switzerland.
I allocated two weeks for this
visit as well, but this time I didn’t want to stay at one location.
I collaborated with Alberto
Lerner from eXascale InfoLab at University
of Fribourg, and several others, recently that led to a CIDR 2023 paper. Alberto
asked me to visit them many times before. So, I reached out to him to make that
happen finally. By the end of my visit, he almost convinced me to go there for
my next sabbatical to learn FPGA programming.
Then, the roots of my academic
family are at EPFL. So, I reached out to my academic mother and Dimitra to visit
EPFL. They kindly aligned the DIAS
end-of-year raclette party with my visit. That party is one of the things I
miss from my PhD years, so I really appreciated this. It was a lovely reunion.
Finally, two of my PhD students (Ties Robroek and Ehsan
Yousefzadeh-Asl-Miandoab) did their research stay abroad (a mandatory part
of PhD at ITU) at ETH (hosted by Ana Klimovic) and University of Basel (hosted
by Florina Ciorba),
respectively. While we unfortunately couldn’t find a suitable date with
Florina, we were able to arrange a visit with Ana. It was my first time
visiting the systems group at ETH. I especially
loved exchanging info and experiences with everyone there working on a variety
of systems topics.
During this leg of the sabbatical,
I also had the chance to see many people that I haven’t seen in such a long
time from academic family to family friends, and I am grateful to my academic
sister Danica Porobic for being my unofficial
host in Switzerland.
All the interactions I had during
these visits were worthwhile on their own even if in the end nothing comes out
of the projects I got involved in or made plans for during this sabbatical. I
am very grateful to my hosts, all the PhD students, and other team members for
taking the time to talk to me.
Practical stuff
Accommodation. Most of my
travel and accommodation expenses during the sabbatical were covered by my own
funding. However, my hosts also helped. Tilmann asked HPI to cover my train
tickets to/from Berlin and gave me really good tips for accommodation, which
helped me to find a very good option with a reasonable price in Berlin (I
booked six months in advance). Peter helped me to get a room at CWI guesthouse,
which reduced my costs substantially in Amsterdam. In general, it is good to reach
out to people early on and ask them for tips for these practical matters.
Thanks to my funding, I was also able
to keep my apartment in Copenhagen as is and was back in Copenhagen a couple of
times in between visits. It was nice to be back home for a bit.
Visas. Since I on purpose
arranged my sabbatical in a way that avoids visa applications, I didn’t have to
worry about this, but acting early would also be important on this matter. On
the other hand, keeping my stays relatively short at different places was
partially motivated by this goal. With my Danish resident card, I am safe
traveling up to 90 days abroad in Schengen region, but not more. Technically,
no one checks or stamps your passport when you are within the Schengen region,
but this isn’t something I would abuse, especially carrying the kind of
passport I carry.
Health insurance. The
health insurance I have thanks to my job becomes invalid for sabbaticals since work
trips that are longer than 28 days are no longer covered. If you are an EU
citizen, you may still be covered through your Blue
Card, but I am not entirely sure, since I am not eligible for this card as
a non-EU citizen in Denmark, so I didn’t investigate it further. Therefore, I
had to find another solution for my health insurance.
Following up on the suggestions
of a few colleagues, I decided to check if I can get a Mastercard
Gold, which provides health and travel insurance for trips up to some
length in Europe, which fit well for my sabbatical case. And since my Danish
bank likes me (the feeling isn’t mutual), they agreed to give me a Mastercard
Gold for a cheap price. Now, I have three cards: one parliament blue, one smoky
gray, and one gold. I wish I had pink, red, and purple instead.
Luckily, I didn’t have to use the
gold card for health insurance purposes, but it was the only option I had while
paying for the long-term accommodation bill in Berlin. The limits of my other
two cards weren’t high enough.
Carrying my stuff. I had
to buy two new carry-on-size suitcases. I won’t go into further details on this
one unless you ask for it, but in general I spent more time than I expected
(and wanted) thinking about suitcases and how to pack them.
Communication. I didn’t
have communication issues except for the French-speaking part of Switzerland. I
got around in Germany easily with the combination of my English, highly
deteriorated German, and native Turkish. In Amsterdam, English was enough, but
I made attempts at parsing written Dutch using all the languages I know. In
Lausanne, I was impressed by my ability to still remember how to order food in
French. It was out of necessity as I didn’t have the option to order things in
English. In Fribourg, asking for directions at the university building in
English got me nowhere, and I texted Alberto.
Other expenses. I never
thought I would say this, but Copenhagen is cheap. I knew it would be cheaper
than Switzerland. But it is also cheaper than Berlin, at least in terms of my
usual university-lunch-plate and grocery shopping. I didn’t stay in Amsterdam
long enough to judge this well, but it felt similar to Copenhagen in terms of
such costs.
Wellbeing
During my first week in Berlin, one
concern a couple of my friends had for me was whether I would feel lonely and
down. Their concerns were valid since I have a track record of being slightly
depressed whenever I move to a new country till I get used to the place (or in
the case of Switzerland, the whole duration). The sabbatical was a
substantially different experience, though. I have neither felt lonely nor
depressed at any point. I would attribute this mainly to two factors: the
social nature of my sabbatical and me being older.
First, my hosts, Tilmann,
Matthias, and Peter, were all extremely supportive and inclusive, and their
groups immediately adopted me as one of their own. There were also a lot of
work-social activities of the good kind, not the draining kind.
The rest of the time, I was going
around visiting different places and giving talks. In a way, I didn’t have time
to be “down”. I had overall constructive interactions and discussions with
people during these visits. There was only one instance where I felt slightly
like shit after my talk because of certain attitudes both during and after the
talk, but other times were all energizing experiences.
Second, I have known myself for
almost 36 years now, and with that knowledge I can better prevent or
damage-control negative experiences. My sabbatical goals #1 and #4, as listed
above, were there because I was trying to avoid certain things that tend to
pull me down. Following those goals, I didn’t need another Western-world
approval for my existence at the places I was visiting, and both Berlin and
Amsterdam are great cities to be in for longer durations for me. In Berlin, my
body felt like she was in her natural habitat. In Amsterdam, I had the same
feeling I had when I visited Copenhagen for the first time (for my ITU job
interview), which I will call peaceful excitement.
On the other hand, my rent-like
period got a bit off during my sabbatical and I had trouble sleeping on some
nights, especially toward the end. Considering my intense and mobile schedule,
this was expected, I guess.
Ending with the most important
point. When I announced my sabbatical, a friend joked “Is this an excuse to
watch Gilmore Girls again?”
I have a tradition of watching
Gilmore Girls every time I move to a new place, which means that I watched the
whole series 5 times and watched parts of it even more times as I introduced it
to other friends. The logic behind this tradition is at a new place, in the
beginning, one must deal with a lot of instability without having much to rely
on. A show like Gilmore Girls gives me the feeling of stability amid the chaos.
My sabbatical stays were somewhat
eligible to put my Gilmore Girls tradition into action. However, shortly before
my sabbatical, I started rewatching Gilmore Girls with my dear colleague Veronika Cheplygina. During my sabbatical, it made sense to pause it and resume with her when I get back. Since I believe
in flexible traditions, keep the core and adjust the content if you need to, I
instead went further back in the nostalgia lane and started to rewatch Buffy the Vampire Slayer. Still
watching it and currently at season 6.
I worked with Erietta
Liarou, who did her PhD at CWI, but I don’t think this counts, since she
was a postdoc at EPFL and no longer at CWI when we collaborated. Also, an unrelated
note, people used to mix-up Erietta and I at conferences a lot.
