Nik Bamert

Hexagon Robotics · Senior ML Engineer2026–

ML systems and inference.

Independent · ML Systems Engineer2025–2026

Work on search, inference and performance-critical C++. Selected write-ups under Projects.

Pallon · Software Engineer, ML / CV Systems 2020–2025

Owned production reliability of the computer-vision pipeline, from 3D reconstruction to inference, as it scaled to millions of images a day. Led ML/CV for a new product line: labeling workflows and tooling, model development, training, inference and deployment. Initiated and carried out the move to explicit contracts between pipeline components, making the system testable end to end and letting the team iterate faster.

ETH Zurich · Research Engineer2019–2020

Continued my MSc work on neural network compression and quantization in the CVG, including custom CUDA/C++ kernels.

ETH Zurich · BSc & MSc, Computational Science and Engineering2014–2019

• MSc thesis: Fully Differentiable Discretized Neural Networks with Shift Convolutions for Embedded Computer Vision. Grade 6.0/6.0, supervised by Martin Oswald and Marc Pollefeys, 2019 (record)
• BSc thesis: Efficient Matching via Hashing for (Close-to) Real-Time Dense 3D Reconstruction. Grade 6.0/6.0, supervised by Torsten Sattler and Marc Pollefeys, 2017 (record)
• Semester project: Two-stage (semi-)dense real-time 3D depth estimation, supervised by Marc Pollefeys, 2018 (record)

• → SIMD (AVX2) implementation of PatchMatch stereo 2019
• → fade: fast adaptive stereo estimation (semester project) 2018
• → openGPC: fast sparse correspondences via hashing (BSc thesis) 2017

Independent · Embedded Systems Engineer2007–2015

Firmware and hardware for industrial clients, among others in public transport. Bare-metal firmware, bootloaders for field updates from external storage media, and a hardware abstraction layer with file-system drivers and drivers for storage media, network controllers and displays. Deployed across Switzerland, wider Europe and the US, including Zurich, and still in active use in 2026.