tebartsch
I love to make computers show intelligent behaviour.
This is the website of Tilmann E. Bartsch. I currently design and implement the
software toolchain of Semron's ai-accelerator. Previously, I studied data
science, mathematics and physics.
If I don't train a neural network or optimize an ai-toolchain you'll most likely find
me in the climbing gym or on the tennis court.
Blog
Quantizing Transformers
The model size and computational costs of modern neural networks are huge. Both can be reduced by using
a method called quantization. This post is a detailed guide on how to quantize neural networks and
builds up everything required to perform a simple quantization of Google's transformer-based
ViT from scratch.
Why you should work on saturdays if you're younger than thirty.
“Knowledge and Productivity are like compound interest” (Richard Hamming). Therefore, the commonplace idea about learning as much as possible when you are young is true in an exponential way.
About
After developing an extensive curiosity in mathematics I started to dive deep
into it at the technical university in the beautiful town of Dresden in Germany.
Over the years as the new machine learning era gained a tremendous amount of
traction I was deeply impressed about the new programs showing something that
is at least on the boundary of intelligent behaviour.
Reflecting on this I developed the aspiration to participate on these advances.
I learned about the necessary steps to obtain the current result in
artificial intelligence ranging from designing semiconductor chips, building
compiler toolchains and training deep neural networks.
Consequently I am very happy to currently have the chance to work on the very
foundations of building next generation ai applications at SEMRON by designing the
software toolchain for our custom inference chip.
CV
Experience
Semron
ML Software Architect
since Apr. 2022
Symate
Software Engineer
Aug. 2021 - Mar. 2022
Education
TU Dresden
Bachelor Physics
Oct. 2018 - Sept. 2021
Bachelor Mathematics
Oct. 2017 - Sept. 2020