What I do as an assistant professor (Fall 2025)

A brief summary of my activity and lessons learned as an assistant professor of computer science at FI MUNI during the fall semester of 2025.

Disclaimer: This report can be interpreted in many ways. If your takeaway is “wow, this guy is working too much” or “wow, this guy is not nearly working enough”, that’s not productive feedback, keep that to yourself :) Also, don’t compare yourself to other people too much. Everybody’s situation is different. With that said, if you do actually have some constructive input on how to improve my time management (i.e., “wow, this guy should really be working on X instead of Y”), I’m all ears.

Foreword

I described my tracking methodology and general objectives in my first article. In case you haven't seen it, a brief summary: I try to objectively track the tasks that I work on during the day, but for the sake of my own sanity, it is only a rough estimate (~15-minute granularity). As such, many smaller tasks get lumped together into larger "catch-all" categories or end up missing completely. The report covers 17.08.2025 - 15.02.2026, i.e., almost exactly 6 months. My expected work distribution is roughly 40/40/20 between teaching, research, and "admin". With ~2 weeks of vacation and 3-4 state holidays, this accounts for 340h of teaching, 340h of research, and 170h of "admin".

In the original article, I also discuss salary a bit. Here, the situation has not changed much, so I will only touch on it briefly. Over the past year, my salary increased to ~88k CZK / month, amounting to ~550 CZK / hour. However, I logged more than 1200 hours of "work" over the past semester, which is 1.4x the "expected value". As such, my "effective salary" is closer to ~390 CZK / hour, or ~63k CZK / month. As we'll see, I managed to do a bit more science this semester compared to the last one, so I guess the "overtime" is somewhat acceptable, but ultimately I was also bogged down quite a bit in various "admin" tasks.

Lessons learned

This semester, I'd like to start with lessons learned, because it is probably the more interesting part in the long term.

  • Theses: I wanted to prepare some "practical guidelines" for thesis writing and advising, because last semester I noticed that a lot of students struggle with writing the thesis text. I did not have time to do this yet, but I want to highlight it here because it is still an ongoing problem. It also affects me as an advisor, because I end up spending a lot of time on repetitive, avoidable tasks. As many have noted, LLMs "help" with this, but are only useful if you can already identify and describe what makes text "high quality".
  • Grading: A similar takeaway could be said about grading and examination in terms of frontend software engineering courses in our program. We have some objective criteria, but they are loosely defined and hard to enforce. We would very much benefit from a document that clearly establishes "a passing grade software engineer must know at least X". Not everything can be formalized, but some baseline is necessary.
  • Student use of LLMs: Because I mostly interact with master's students, my view of LLMs is perhaps a bit unexpected: I believe students are not using LLMs (well) enough. At the moment, most students can "chat" with an LLM, and they can paste an assignment into an LLM to get some passable solution, but (so far) they very rarely use LLMs to do actual quality control, automation, or complex planning (e.g., despite LLMs being very popular, we still get a ton of submissions with trivial mistakes that a "review this" prompt would identify immediately). Some students use LLMs reasonably well for complex, multi-step coding tasks, but this is still a minority. In my opinion, more challenging and comprehensive assignments are needed, but this also complicates other steps of the teaching process.
  • My personal LLM workflow: Since spring, I use LLMs heavily for coding, but my work is mostly divided into two types of tasks now: Mission-critical code, which I still write and test mostly manually, and various "auxiliary" tasks that can be safely delegated and don't require in-depth understanding (or have a simple but exact specification that the LLM can access). Here, I find that frontier models can be helpful, but are unnecessary for most tasks, assuming your planning and validation loop is robust (and your project isn't a mess). Where frontier models with grounding are absolutely necessary is any kind of "deep research". Here, smaller models hallucinate too much. Large reasoning models can catch many hallucinations via introspection and by comparing with real-world references. I also use LLMs to review most of the text I publish. LLMs can now (mostly) update text while preserving my personal style. Furthermore, LLMs are now very good at checking non-trivial math and proofs. They are not perfect (review is still necessary), but they catch issues that I would normally spend much more time on.
  • My (somewhat depressing) takeaway is that on average, spending 20-40h advising a student on their bachelor's or master's thesis is mostly pointless, unless that student has non-trivial potential to become a contributor in your lab, or a tutor in your course. In any other situation, you can get more, higher-quality output by spending that same amount of time working with a good LLM. That does not mean tutoring students is pointless now, it just means the previous value proposition of "I supervise a topic and the student helps me explore it" is no longer there, unless the student is significantly above average, or has potential to contribute in other aspects beyond the thesis.
  • Another negative observation is that given my current occupation, I practically no longer read or write for fun. I still code "for fun" a little bit, but the idea of "relaxing by reading" is now quite alien to me.
  • Overall, I think that in the current setting, I am starting to experience the symptoms of a "manager's brain": While I am still contributing individually to many things, there simply isn't enough time to focus on any single task properly. In the long term, this negatively impacts the quality of my work and my cognitive abilities.

Time analysis

This semester includes ~15 work days "on vacation", because it covers a large part of my summer holiday as well as Christmas. It also includes 4 national holidays and 4 sick days. As we'll see, this does not really affect the overall calculation that much, though.

For anyone impatient, here is first a breakdown for this semester. Comparison with the previous semester is at the end of the article.

Visualized: Semester Time Breakdown

Administration

It is quite complicated to classify specific tasks as “administrative”. My own classification mostly assumes that a task is administrative if the output is not a useful “deliverable” in the teaching or science categories.

Teaching/faculty administration (~202h)

Administrative tasks include:

  • (136h) “Emails” (see below)
  • (20h) Various planning and overhead, including this report
  • (17h) Administrative tasks (teaching agreements, travel reports, ...)
  • (12h) Faculty-wide and department meetings
  • (9h) “One-on-one” meetings; typically miscellaneous planning or consults
  • (8h) Outreach (open days, lab day, ...)

Emails: This includes reading and responding to emails and messages, as well as doing “quick tasks” triggered by a message. In particular, this includes a lot of administrative tasks that are not logged separately. Bottom line: Strictly speaking, not all of this time is spent responding to emails, but it typically corresponds to time spent on communication and miscellaneous reporting.

We can also quantify this. As discussed last time, I have a relatively good idea regarding how many "notifications" I get via email (those end up in the trash) and how many "conversations" I interact with (something that I do actually respond to). Compared to last semester, the number has increased slightly. I generally delete 1000-1500 emails per month, and react to 300-350 conversations (worst months are closer to 450, though). This also does not account for direct messages. Overall, this is consistent with the increase in time tracked in this category (last semester, it was 103h), where we estimated that each deleted notification email takes 2-3s to process, and each conversation takes 3-4 minutes to resolve. However, I am suspecting the error bars are very large on this, so this is only a "napkin math" estimate to figure out if this semester is an outlier.

Research group administration (~165h)

These tasks mostly cover things related to infrastructure and administration within our research group. I’m not putting these in the “science” category because they do not contribute directly to specific papers/grants, but are nevertheless necessary to maintain our scientific output.

  • (42h) Maintaining lab "Biodivine" codebase
  • (28h) Maintaining BBM model repository
  • (28h) New low-level utility libraries for Biodivine
  • (28h) Other infrastructure maintenance and experiments
  • (20h) Social and planning meetings within the lab
  • (15h) Lab seminar
  • (13h) Lab-related travel

Compared to the previous semester, this item has grown significantly, especially in terms of maintenance and coding. While not ideal, I believe these elements will effectively contribute towards better scientific output in the future, hence I don't mind this increase that much. Furthermore, these are usually tasks that are hard to delegate, because they require a disproportionate amount of onboarding. Compared to last semester, the social and community aspect is a bit lower (something we definitely need to work on).

Summary

Overall, I spent 367 hours performing administrative tasks and maintenance compared to 170 “planned” hours (more than double). This is a major increase compared to the last semester, but at least in the lab category, it should produce justifiable results. Emails and communication are still a major category (~136h). I don’t think this part can be significantly simplified without major changes in my duties.

Teaching and advising

Advising and reviewing bachelor's and master's theses (~134h)

I interacted with 16 students as a thesis advisor or in some form of consultant position. Out of those, 3 defended their theses this semester. I also reviewed 4 other theses.

I was the consultant for two PhD students. Ideally, I would count this as research, but our evaluation scheme considers this to be teaching, so I included it here :)

  • (76h) Meetings and reviewing the work of students.
  • (26h) Thesis-related overhead, like topic materials (7h), defense (6h), and writing official reviews (13h).
  • (32h) Meetings and material preparation for PhD students.

Obviously, this was a significant portion of my teaching duties this semester, but is also significantly lower than last semester. This is partially due to a smaller number of students defending this semester, partially due to better time management, but also (likely) a lower level of engagement on my part.

Courses and teaching materials (~265h)

In the fall semester, I was involved in the organization of the following courses:

  • PV252 (Web Development), ~75 students, 3 seminar groups. I am the main lecturer, and I also ran two seminar groups this semester. This course is also currently "under development", so it requires a lot of new materials.
  • IB111 (Introduction to programming), ~20 students, 1 seminar group. I was running one seminar in this course, but I am otherwise not involved in its organization.
  • PV247 (Modern Development of User Interfaces), ~50 students, 2 seminar groups. I am the faculty member responsible for this course, but the actual materials are prepared and taught by external contractors.
  • PV292 (App development in Flutter), ~70 students, 3 seminar groups. I am the faculty member responsible for this course, but the actual materials are prepared and taught by external contractors.

  • [IB111] (24h) Teaching at the IB111 seminar.

  • [IB111] (15h) Reviewing student homework assignments and preparing discussion topics for the seminar.

  • [PV252] (78h) Preparation of new materials (several new or significantly updated lectures and seminars).

  • [PV252] (56h) Teaching the UI segments in two seminar groups and most course-wide lectures.

  • [PV252] (38h) Project grading, interviews, and presentations.

  • [PV252] (19h) Course administration (projects, assignments, planning, ...).

  • [PV252] (16h) Homework assignment grading.

  • (19h) Activities related to other courses (mostly preparation for the spring semester).

Overall, this is more teaching than last semester, primarily due to material preparation. So hopefully, this should get easier next year.

Summary

Overall, I spent 399 hours performing "teaching duties" compared to 340 “planned” hours. This is less than the previous semester, but most of that time seems to have moved into "admin" tasks.

Science (~447h)

Here, task classification is also somewhat complicated, because we can split the tasks either by project or by the type of activity (meeting/discussion, writing, coding/experiments, …). First, let’s list tasks that don’t belong to any specific “project”:

  • (44h) Writing and refining a grant proposal.
  • (35h) Attending the CMSB conference.
  • (34h) Various exploratory experiments, learning new tools, ...
  • (13h) Reviewing several large journal papers.
  • (7h) Preparation of various talks.

Now let’s list the other “project-oriented” tasks, with the caveat that some stuff is not published yet, so I’ll need to be a bit vague about it. Another issue is that for these tasks, I often don’t have a clear distinction between meetings and focused work (plus some meetings fall in the "consulting PhD students" category), so we will again need to accept some reasonable estimates. Also note that most of the larger projects span multiple semesters, so we can’t get a complete “cost” of any project based on these numbers (we should be able to get some better numbers eventually, but that will take a few years).

Summary

Note that some items (BBM repository) that I previously counted as research have now moved to "admin", yet this category grew significantly compared to last semester, and is probably the main contributor towards my (lack of) work-life balance. Nevertheless, the results do seem promising, so at least for now, it seems the effort is paying off. Also, as expected, projects with many collaborators have a lot of overhead compared to smaller, more focused tasks.

Wrapping up

The first important observation is that last semester, my average work week had one extra day (~48h instead of 40h). Now, it is 58h instead of 40h, so more than two extra days. This may be somewhat influenced by the fall semester including most of summer, but is overall not a good trend (and I don't see a major improvement in spring 2026 so far). As such, next steps will definitely focus on trying to get this number into acceptable levels. However, another important observation is that so far, the extra work is mostly paying off in terms of new publications and collaborations. Nevertheless, I still need to learn how to delegate better, and more importantly, find people to whom I can delegate.

Semester-over-Semester: Time Spent by Category