Mattias Stahre

Senior AI Engineer who gets genuinely excited about making machines think better 🤖

Currently: Building magical learning experiences at Sana using LLMs that actually understand how people learn. The kind of AI that makes you go “how did it know that?”

Previously: Made radar signals dance at Saab. Taught neural networks to see patterns in defense systems that humans couldn't spot. Built tools that engineers actually wanted to use.

The messy, beautiful intersection of research and production code.

Side note: Also won some racing championships 🏁 and made commercials for IKEA 🎬. Turns out pattern recognition works everywhere.

Published:

• Data Collection and Layout Analysis on Visually Rich Documents using Multi-Modular Deep Learning (Master's Thesis, KTH 2022). Evaluated LayoutLMv2 for document understanding and built a pipeline for VRD datasets.
• Deinterleaving of Radar Pulses with Batch Processing to Utilize Parallelism (Bachelor's Thesis, Saab/KTH 2020). Evaluated batch processing for parallelism in radar signal deinterleaving.

Fun fact: I speak fluent Python, decent Mandarin, and terrible Swedish jokes.

mattias@stahre.nu | LinkedIn | Stockholm (the good one)

p.s. If you're building something that could change the world, let's talk.