About

I’m a machine learning engineer based in Zurich, working on AI evals, risk, and safety.

2025 –  Engineering Manager, LatticeFlow AI

I built the first prototype of LatticeFlow’s AI governance platform from scratch and owned its architecture. It started as an MVP with a team of three. The company then pivoted in that direction and eventually the whole team moved onto it. This led to multiple enterprise clients and demos.

My focus is the evaluation layer: 100+ evaluations for things like jailbreaks, prompt injection, bias, harmful content, hallucination and data leakage, mapped to governance frameworks and standards like OWASP and MITRE. It also does custom metrics, model-as-a-judge scoring, Pass@K and repeatability checks, and dataset generation. A lot of the work is plumbing: caching, rate limits, parallel runs, refusals and retries. Every run keeps a record of how it was produced, so a governance claim can be traced back to it. I also worked on COMPL-AI, a technical interpretation of the EU AI Act.

On the management side, I led a team of up to 8 across 3 locations.

2023 – 2025  Senior Machine Learning Engineer, LatticeFlow AI

Mostly data quality at scale: finding duplicates, mislabeled samples and blind spots with embeddings, in workflows that ran over a million samples at a time. I also worked on interpretability and failure analysis methods (GradCAM, attention rollout, feature-based blind spot detection) to debug model behaviour. For one automotive customer this raised car segmentation accuracy by 40% and damage detection by 10%, mostly by fixing the data rather than the model.

2021 – 2023  Machine Learning Engineer, LatticeFlow AI

I joined as employee #6, on the core ML team during the company’s first product phase. I led the work to bring object detection into the platform, writing model wrappers and feature extractors for models like SSD, Faster R-CNN and YOLO. I also built core platform features such as data versioning, incremental upload and incremental computation. Beyond that, I worked a lot on POCs and client engagements.

2019 – 2021

M.Sc. in Electrical Engineering and Information Technology at ETH Zurich, graduating with honors (5.8/6). My focus was machine learning and computer security - which turned out to be a pretty useful combination. Along the way I published work on adversarial robustness at NeurIPS 2022 and IEEE CyCon 2024.

2015 – 2018

B.Sc. in Electrical Engineering and Information Technology at ETH Zurich, focus in signal processing. This is where I first got pulled into deep learning and never quite left. In between degrees I spent six months as a software engineering intern at Open Systems and traveling the world.

What I’m interested in

AI safety, security and interpretability.

Outside of work I like being outdoors: kitesurfing, ski touring, hiking, etc.

Find me on GitHub or LinkedIn, or reach me at: