Markus Göckeritz

Zurich, Switzerland

About Me

Seasoned fullstack software engineer, team lead, and product manager with 10+ years background in creating and executing innovative software solutions. Highly experienced on the entire software stack from idea through to development, deployment, and maintenance. Consistently recognized as a core contributor and competent leader.

I am incredibly passionate about software and care for the tiniest details in the products that I build and ship. I take pride in well thought-out and carefully crafted solutions and I enjoy solving complex issues.

Work Experience

sixa AG Sep 2019 - Jul 2022

Co-Founder
Head of Engineering / Director

sixa AG is an online marketing agency with now 20+ employees that I co-founded with two friends. Next to my managerial duties as a founder and director of sixa, I mainly focused on the technological aspects of the company, client projects, and internal projects.

For clients, I developed all sorts of software ranging from API and system integrations to web apps and full-fledged online stores.

In parallel, my main focus was on developing WordPress extensions as part of an isolation strategy to improve the efficiency and effectiveness of our work, but also to be offered as standalone products in the market. So far, we have developed 10+ publicly available libraries, 5+ templates, and 45+ plugins that are used accross various projects.

curv.io Jan 2018 - Sep 2019

Co-Founder
Lead Software Engineer / Product Manager

Curv.io was a web application for SEO professionals. Key features included keyword research, a keyword tracking dashboard, and a competitor overview for large volumes of keywords. Curv.io was discontinued in 2019.

My responsibilities included:

Integration Alpha GmbH Apr 2015 - Aug 2019

Software Engineer

At Integration Alpha, I gained insights in many different roles and various projects and industries.

I joined Integration Alpha as a software engineer and developed a natural language processing pipeline in Java during my first year. In my second year, I was working as an IT consultant on big data projects for major financial institutions in Switzerland. After this, I was working as a software engineer and product owner on a web app for container logistics where I led a team of 4 developers.

University of Zurich Feb 2018 - Jun 2018

Teaching Assistant, Software Engineering

Software Engineering is an advanced undergraduate class at the University of Zurich in which students are assembled into small groups and tasked to build an application. Usually, this application is a game of some sort (back in my days we built an Android game). During the spring semester of 2018, the application to be built was a browser-based game of El Dorado. The technologies used were ReactJS and Java.

My duties included mentoring, helping, and advising students as well as grading their assignments. During the course of the class, I have supported my students in building the game but also in engineering the technical requirements, documenting the software and specifying the architecture (following the guidelines of the class).

SWIT Solutions AG Feb 2015 - Jun 2015

Software Engineer

One of the core products of SWIT Solutions AG is EuroTime, a personnel planning and time recording application. The application was built with C# / .NET. As part of a university project, I was tasked with developing a browser client for EuroTime. Together with a good friend and fellow student, we successfully delivered an MVP browser client within three months.

University of Zurich Feb 2015 - Jun 2015

Teaching Assistant, Software Modeling

During the spring semester of 2015 (my 4th semester), I was a teaching assistant in an undergraduate computer science class at the University of Zurich. My duties included holding biweekly lectures and grading student assignments. The class was about software modeling and requirements engineering and we covered the following topics:

Education

University of Zurich Sep 2017 - Mar 2020

Master of Science in Informatics
Data Science
summa cum laude

Thesis: Eager Machine Translation

Eager machine translation is a novel approach to simultaneous machine translation that produces exactly one translation token for every input token. The model is trained with a carefully pre-processed training set that is supposed to allow the model to learn to produce WAIT tokens (to be able to exhibit a waiting behavior).

Originally, the eager translation model uses beam search to improve the quality of the translations. However, beam search can only be applied subsequently once the entire sentence is translated, which defeats the original purpose of translating simultaneously.

Research has shown that beam search can be eliminated by applying knowledge distillation. Knowledge distillation is a process where a so called student model is trained with a training set that is produced by another model, called the teacher model.

In my thesis, we extended the original eager machine translation model and analyzed the model quantitatively and qualitatively. We have shown that knowledge distillation does indeed remove the need for beam search and demonstrated that the performance of the model can be increased substantially by using a transformer network as the teacher model (although the teacher model still outperforms the student model by a substantial margin).

Further, we investigated the effects of knowledge distillation by analyzing the monotonicity, parallelism, and determinism in all outputs. We confirm the tendency that these three measures increase through knowledge distillation. We were unable to identify the relationship between the magnitude with which monotonicity, parallelism, and determinism increase through knowledge distillation and the improvement in the translation quality.

University of Zurich Sep 2013 - Dec 2016

Bachelor of Science in Informatics
Business Informatics
summa cum laude

Thesis: Quantifying and Correcting the Majority Illusion

Under certain naturally occurring configurations of a social network, an attribute may appear popular among local clusters, even though it is globally rare. Similarly, if a large number of local clusters misperceives the popularity of an attribute, the attribute may falsely appear popular on a global scale. This network phenomenon is known as the majority illusion.

In social networks, opinions and behaviors spread similar to communicable diseases. Thus, the perceived popularity of an opinion or behavior substantially contributes to the spread of said opinion or behavior (oftentimes referred to as social contagion).

In my thesis, we developed algorithms to detect and quantify the majority illusion in social networks, both on a local and a global scale. Additionally, we developed an algorithm that causes the majority illusion and benchmarked it against a naive approach using simple network heuristics in different common network structures. We have shown that our algorithm is capable of estimating target vertex sets with comparably small effort on a variety of generated networks with different network characteristics. Lastly, we introduced an alternative dynamic diffusion model to simulate the spread of a social behavior in a network of interacting individuals.

Awards

University of Zurich April 2017

Semester Award

Each term, the University of Zurich awards 30 outstanding term projects and theses from all faculties with the Semester Award. I'm honored that the University of Zurich has awarded me with the Semester Award for my Bachelor Thesis Quantifying and Correcting the Majority Illusion.

Languages

German: Fluent

English: Fluent

Swiss German: Fluent

French: Juste un petit peu.