Hi👋, I’m Fabian, a data scientist bridging machine learning and engineering.
Scroll down and join me on my journey.
Fabian Rosenthal
Data Scientist
With a M.Sc. in Media Technology, I specialize in developing robust machine learning solutions that perform reliably across simple and critical applications.
Spezialization
My research on uncertainty quantification and cross-validation has equipped me with deep insights into performance estimation and statistical computing. This allows me to make data-driven decisions.
Projects
In my projects I like to incorporate my knowledge to build reliable solutions.
Projects
In this forecasting of bicycle demand in Cologne, I incorporated prediciton intervals to account for the uncertainty in the model. This gives more insights to the stakeholders and helps to make better decisions.
Master thesis
In my thesis at Fraunhofer MEVIS I trained 4+ million models to find appropriate methods to get confidence intervals for cross-validation estimates.
Master thesis
The complex study design allowed me to analyse the operation characteristics of statistical methods and to derive recommendations for the application in practice.
Programming
I loved the development of my package for complex cross-validation workflows: flexcv. It uses a class interface for setting up the experiments.
Work
flexcv originated in my work at Hochschule Düsseldorf where I cross-validated a lot.
Work
In my 3 years as a student research assistant I developed the ML pipelines that are used in a study on sound perception in private dwellings. I was also responsible for the processing and feature engineering of the 6000+ audio recordings of our study.
Work and Bachelors
A comprehensive study on feature selection including 2000+ audio features was part of my Bachelor thesis. I presented my work at DAGA 2022 in Stuttgart.
Background
During my studies in audio and video engineering, I transitioned to a more data focused path.
Background
Learning about machine learning in a project on music recommender systems at Hochschule Düsseldorf was a game changer for me and motivated me to pursue a career in data science.
Background
And on my journey I mastered a lot of languages, frameworks and tools.
Background
My interdisciplinary background—spanning audio engineering and data science—enables me to approach complex challenges with a holistic, end-to-end perspective. I’m passionate about working in cross-functional teams and collaborating with domain experts to deliver impactful solutions.
Let \(W\) represent my work model \[ W = \tau \cdot m \cdot ( b + P_i \cdot \gamma ) \\ \]
Objective: \(\max_{m, \gamma,P}\)
Where:
- \(m\): Motivation with \(m \sim \mathcal{U}[0.7, 1.0]\)
- \(b\): Intercept denoting the engineering background
- \(P_i\): Matrix of \(i\) personal trait properties
- \(\gamma\): Vector of learning rates corresponding to \(P_i\)
- \(\tau\): Team spirit