Michał Wysocki

I build machine learning systems that model and predict the real world.

Autonomous systems and multimodal detection in practice; world models and representation learning in research.

Currently

Junior ML Engineer

Sep 2026 — present

Hertz New Technologies

Machine learning for autonomous systems and counter-UAS: multimodal detection, computer vision, prediction, and sensor data.

Experience and education

Selected Work

Anchor

In progress

path consistency for latent world models

A self-supervised mechanism that penalizes disagreement between horizon-conditioned direct prediction and autoregressive rollout, aimed at reducing compound error in latent rollouts. Designed as an importable component rather than a standalone architecture.

Codeworld models · self-supervised learning

acoustic early warning, and where it breaks

A Random Forest classifier on MFCC features with local inference, reaching strong held-out metrics that then failed on real recordings while misflagging ambient noise. The generalization gap, not the accuracy number, is the finding.

Codeaudio ML · sim-to-real · embedded

All projects