Hi, I’m Jakub (Kuba) Strawa 👨🏻💻
I'm AI researcher (ex-AI engineer) focusing on multimodal models, post-training, alignment, and reinforcement learning. I’ve built and served AI applications for Fortune 500 companies, worked in R&D at Raiffeisen Bank on custom multimodal solutions, and I’m currently an AI Engineer at Roche. In my spare time i share my knowledge in technical posts about AI on my blog.
Research interests
- Multimodal models (vision, language, audio)
- Post-training: preference learning, distillation, and evals
- Alignment & safety
- Reinforcement learning
Latest publication
- Improving KAN with CDF normalization to quantiles , J. Strawa, J. Duda
Elsewhere