ljstroemsdoerfer

View JSON Representation

Head of Data Science @Mercedes-Benz

Funding Links: https://github.com/sponsors/LJStroemsdoerfer

• Name: Lukas Jan Stroemsdoerfer
• Kind: user
• Followers: 11
• Following: 4
• Total stars: 14
• Repositories count: 17
• Created at: 2022-11-15T21:32:00.475Z
• Updated at: 2025-03-09T09:11:19.889Z
• Last synced at: 2025-03-09T09:11:19.889Z

GitHub Sponsors Profile

Production-ready Data Science
Well, there isn't really much to say. However, if you check my bio, you will see that I am most interested in machine learning deployment. Let me tell you why:
I've been working as a Data Scientist for quite some time. At the beginning, most projects demanded some kind of Proof of Concept (PoC). There we would demonstrate what models could do, how nicely they could have predicted some previous shifts or anomalies in the data. This was all fine and fun, however, at some point, PoCs needed to be transformed into a productive piece of software.
Productionizing is a whole of another story than just building a fancy nice PoC with a premium notebook. Nevertheless, I went down this road and quickly discovered, that for machine learning systems, there is little consensus on how to deploy or productionize what and where.
After a while a lot of "platforms" started to pop-up claiming to ease development and production. Some of them are working great and definitely earned their spot on the market, don't get me wrong. However, they are often pricy and force you to leave your beloved IDE and comply with a bunch of rules or guidelines specific to the ecosystem of the "platform" you are using.
I've worked with numerous platforms, however, I always felt the unfulfilled urge to stay within my habits, yet allow my code to be agnostic of any kind of "platform" technology. Obviously, I turned to container technologies and their orchestration.
However, these tools are not straightforward and you will always have to leave your beloved IDE or programming language, or even setup entirely new environments on your machine or server. Those will, of course take some maintenance.
Thus, my mission was born: production-ready Data Science without unnecessary complexity.

• Current Sponsors: 0
• Past Sponsors: 0
• Total Sponsors: 0
• Minimum Sponsorship: $5.00

Featured Works