My name is Paul Grillenberger (GriP for short) and I’m a mathematician working as a programmer pursuing music composition as a hobby.
While I concentrated on learning applied maths through stochastics during my bachelor’s, I went on to learn more about the pure side of maths during my master’s (differential geometry) and while earning my doctoral degree (algebraic topology). While the latter two phases enabled me to think abstractly, to generalize solutions quickly and to see connections between seemingly unrelated areas, the first phase accelerates my learning process in Data Science, especially related to Machine Learning.
At the moment, I’m learning about the foundations of Machine Learning, i.e., how to model the learning process mathematically, how to evaluate whether a certain Machine Learning model might be superior over another in certain circumstances and whether a certain problem is learnable at all. One point that’s most important is that Machine Learning is not the Holy Grail that solves and automates every problem in existence. Machine Learning is most useful when we’re dealing with a problem that can be solved by experts, given a sufficient amount of time to evaluate the data. If the evaluation is tedious and takes a quite long time, Machine Learning might come in handy to accelerate the process. Even so, it is obligatory to utilize a priori information about the problem; a Machine Learning model trained with generic data without any assumptions can only give very generic answers.
Links and resources
I’m using the following resources:
- Hastie et. al. – The Elements of Statistical Learning
- Bishop – Pattern Recognition and Machine Learning
- Shai and Shai – Understanding Machine Learning: From Theory to Algorithms
The latter can be downloaded for free via their linked web page.
About my thesis
In case you want to take a look at my thesis, see below.
I have been working as a software developer from October 2016 to July 2020 at Collenda GmbH, where I have used Java and Oracle PL/SQL as my primary programming languages. Additionally, I have worked on prototyping some Machine Learning models and some ETL stuff with Python which I prefer to use as a language. Since August 2020, I have been working for nexible GmbH. Programming has always fascinated me as it is the ideal companion to the more abstract mathematical world; how I wished I could turn on a debugger and change the values of certain variables to see how it turns out when I was wading through an intriguing but intransparent proof deep inside an article during my graduate studies! In fact, reading and changing code is very much like reading an article and trying to apply its results to the problem you’re currently working on. In this metaphor, test driven development is quite similar to the axiomatic approach I like to take when doing mathematics: You don’t immediately give an intransparent formula and work your way through proofs by crunching numbers and equations, but you pin down the key aspects and then construct a certain object that satisfies those. Extracting such key aspects is as challenging as writing good tests because each test should pin down one aspect of the feature you are trying to implement.
At the moment, I’m learning about the Spring Boot framework.
Links and resources
- Text editors: Vim and Emacs with Evil mode
- Java IDE: IntelliJ
- PL/SQL: SQL Developer
- Python IDE: Visual Studio Code
- I prefer to use zsh with a fork of zprezto as my shell
For more details about my setup, take a look at my dotfiles.
I have been writing songs since I was 14; being a self-taught guitarist and singer, my musical journey started when I still went to primary school and took E-Organ lessons for a few years. This enabled me to learn keyboard and, finally, piano on my own. In recent years, I moved away from using the guitar as my primary composition instrument and like to use the piano as the central piece in my songs. I’ve also started to rearrange some old songs of mine on the piano: Letting Go is a good example.
Links and resources
If you want to listen to a few songs of mine, you may do so on my soundcloud page.
Setup for recording
- Software: Logic Pro X
- Hardware: Some 2012 iMac and the mixer / audio interface Yamaha AG 03.
- Vocals: An old Fame C010 recording microphone that I seek to replace in the near future.