Work

2005-2010: VZLU Prague, research engineer

I started working here during the MSc studies. My job consisted in developing and applying software for optimizing airfoil and, in particular, wing shapes for selected optimization criteria (lift, drag) and constraints (span, depth etc) using an in-house micro-genetic algorithm. I used a lot of these results for my PhD thesis.

We used XFoil as an established fast nonlinear viscous flow solver and coupled it with the in-house μ-ARMOGA micro-genetic algorithm using bridge code written in Python. For whole wing optimization, XFoil was coupled with GNU Octave as a nonlinear equation system solver and the system was again hooked through Python glue code into the micro-genetic algorithm.

 As part of this job, I heavily contributed to the GNU Octave project, an open-source free alternative to Matlab, improving a lot of its aspects, but especially focusing on performance optimizations.  The majority of the performance optimizations in version 3.2 and 3.4 were my work, the most impactful probably being the indexing rewrite (using polymorphic foldable index objects) and lazy contiguous subranges

For the purpose of iterative dense nolinear equation systems solving/optimization in GNU Octave, I developed  the standalone qrupdate library (Sourceforge, GitHub fork) which remains to this day likely the most comprehensive freely available library on the web for low-rank matrix factorization updates, supporting various low-rank updates (rank-1 matrix addition/subtraction, column/row insertion/deletion/rotation)  to the QR factorization, Cholesky factorization, and LU factorization.

This job was a lot of fun and exploration. I remain grateful to VZLU and to GNU Octave for the opportunity. 

2010-2016: Google Switzerland, site reliability engineer

The SRE work at Google consisted, roughly speaking, 50% of incident response for major services (so-called "carrying the pager") and 50% of engineering more reliability into said services in order to keep the amount of incidents at bay. The team was split between Switzerland and California, and had a 24x7 5-minute initial response SLO. There was certainly an element of adventure to it.

I was responsible for core distributed backend services handling account data, account authentication, login, account settings, contacts and profiles. During my SRE time the services grew to tens of millions of requests per second.

I was involved in addressing several high-profile outages and security incidents. Details are, of course, confidential. 

I reached the rank of Senior Site Reliability Engineer during this time, securing two promotions. At the end I served as the Tech Lead for a team of 6 people.

2016-2023: Google Germany, staff software engineer

In 2016, I relocated within Google to Munich, transitioning to a "regular" software engineer role (software development).  I stayed in the internal account and authentication infrastructure area, working to some extent on the backend services I used to support as SRE, but focusing primarily on client libraries that other Google services used to call the account & authentication backends. Literally almost every authenticated request to google.com went through layers of my code, often multiple times.

I later expanded my work to also internal authorization and privileged access management libraries, protecting services against potential malicious or hacked insider accounts.  I called this sometimes "protecting Google against Google".

I reached the rank of Senior Staff Software Engineer during this time, achieving two more promotions. At the end I served as the Tech Lead  for a team of 10 people.

2024+: Google Czech Republic, staff software engineer

After my 13-years Swiss-German corporate ride, Google allowed me to relocate back to Czechia, keeping me in a remote software engineering role.  I look forward to new challenges. Still working on accounts and authentication, but it's different. But it's still the same. 

Miscellaneous