Suggestions
Mateusz B.
Principal Software Engineer @ NVIDIA | MSc in Computer Science
Meet Mateusz Berezecki, a seasoned software engineer with diverse experience across various industries. Holding an MSc degree in Computer Science from Politechnika Wrocławska, Mateusz has made his mark in the tech industry having worked with several prestigious organizations, including Facebook, Two Sigma Investments, Digg, Opera Software ASA, and Optimization Services, LLC, to name a few.
Mateusz's expertise extends to troubleshooting, and he holds a strong passion for helping others with their tech-related issues. He served as a former Software Engineer at Building 8 at Facebook and was also a project manager and technical lead at Sound Control Media Protection. He has also served in other roles, including individual consultant at Devicescape Inc. and software developer at Opera Software ASA. His diverse background has equipped him with unique insights into the tech landscape, making him a valuable resource to anyone seeking his services.
With Mateusz on board, you can expect top-notch service delivery in software engineering, operations research, project management, and more. His broad range of experience and expertise sets him apart, making him the go-to person for organizations that want top-quality solutions to their tech problems. Contact Mateusz today and experience his top-notch services firsthand.
Highlights
I grew up in Soviet and post-Soviet Poland and as part of our education in the civics classes in the elementary and high schools, we analyzed newspaper articles and headlines to learn about slant, bias and attempts of manipulation. It was a useful practice for acquiring and building critical thinking skills. The idea was to not take things on face value, because the Soviet system was really good at manipulating the public opinion.
It is a habit that doesn't go away easily. You can imagine my joy when I get to apply the same skill, to a different type of manipulation; one that I flippantly call the market based propaganda.
With so many people concerned about AI based misinformation, I present to you an exhibit A, written by humans.
Who is right? The answer is to always look at the data.
But what is a poor LLM to do? Who does it trust?
After 10 or so years of waiting for GPU accelerated library for data visualizations to materialize (looking at you matplotlib ;-)) I’ve decided to take matters in my own hands. Because I care about scalability and performance I was curious as to what is possible in terms of data visualization when GPU is involved. To make things more challenging and interesting I started with an iPhone for a more resource constrained environment. Below is the iOS running a scatter plot of some dummy random data with 1,000,000 points, alpha blending for nice transparency effects and real-time rendering. It’s pretty smooth and can deal with more than 10M points easily supporting full interactivity.
The design is simple and consists of a vertex and a fragment shaders, a set of 3D transform matrices and the data itself. Everything renders to a texture that is updated interactively. A design that is familiar to all game developers and graphics programmers.
If you have a wild software idea of your own, give it a try. The key is to experiment. Not every software has to be glued together from existing off the shelf blocks. Sometimes the experiments can yield insights that the old limitations and paradigms no longer apply, because the hardware has gotten so much better. Leverage it if you can.
#makesoftwarefunagain
