Nik Bamert
ML Engineer based in Zurich, Switzerland.
I build efficient ML systems where algorithms meet production constraints.
My background spans production CV/ML systems at Pallon, where I owned production operations for 3D reconstruction and inference pipelines processing millions of images per day; research on efficient representations for computer vision at ETH Zurich CVG, including feature matching and quantized neural networks; and independent consulting on resource-constrained embedded software in public transit systems.
You can reach me through LinkedIn or GitHub.
Featured
Building a Semantic Search Engine for arXiv with Fine-Tuned Embeddings
AgenticKernel: 43× C++ Kernel Speedup with LLMs
Speeding Up Binary Feature Matching with a Cache-Friendly Hash Join
35-Class Audio Keyword Spotting on STM32 ↗
SIMD Implementation of PatchMatch Stereo ↗
Fast Adaptive Depth Estimation ↗