CareerPlan

Master Thesis Title: Learning-based World Representation

ZurichFull-time2026-11-12 → 2026-11-19
On-site

Core

Investigate encoding perception sensor data streams (e.g. RGB images) into compact latent representations and reconstructing features like geometry or semantics.

Role type

Master Thesis Researcher (Robotics)

Builds

Compact latent representations of perception data for robotics tasks

Domain

Robotics, Computer Vision, Machine Learning

Deliverable

research

Required skills

Machine Learning, Computer Vision, Latent Representation, Sensor Data Processing, Robotics

Preferred skills

Deep Learning, Sequence Models, Geometry Reconstruction

Technologies

RGB images, Perception sensors, Latent spaces