About

I’m a Master’s student in Artificial Intelligence at Vrije Universiteit Amsterdam. I’m working with Shujian Yu on my master thesis about utilizng GNNs, CNNs and Information Bottleneck for multi-modal models in the medical field.

Previously I worked at Huawei for around 2 years on deep learning models for Automatic Speech Recognition and later as a full time employee on efficient concurrent algorithms (in China!) Currently I’m exploring different branches of AI to see what’s out there.

If you’re interested in working together feel free to drop me an email at fjakub.off@gmail.com!

Currently I’m interested in utilizing Information Bottleneck (see here) for beter generalizability and explainability of DNNs. I’d also like to work on representation learning, especially for multimodal data - there’s a lot to discover in this area from information-theoretical point of view.

Something I’d love to explore but never really found time is improving efficiency of already existing architectures - especially low level stuff.

You can find example of my academic writing in the Projects section.