Foundation Models Research Scientist: World & Site Memory

Laelaps AI Ltd.; Zürich; Presencial; CHF 120.000 - 150.000; Jornada completa

Descripción de la vacante

Laelaps AI Ltd. in Zurich is seeking a research scientist to lead work on world models, self-supervised learning, and probabilistic spatio-temporal modelling for our LEA foundation model of the physical world.

You will advance site memory, anomaly detection, and multi-sensor data integration in production-ready systems. You will collaborate with PhD-level founders, design experiments, and decide what ships.

Formación

• PhD in machine learning, computer vision, robotics or related field with first-author publications.
• Experience with world models, video or latent prediction, self-supervised representation learning.
• Experience with anomaly or novelty detection where positives are rare.
• Strong PyTorch and the engineering to run your own experiments at scale.

Responsabilidades

• Site memory: a learned, probabilistic model of each site capturing people, vehicles and robots by hour and weather.
• Anomaly as surprise: replace rule-based alerting with measured surprise and evaluate rare incidents.
• World model: post-train open world-model weights on aligned multi-sensor sequences, predicting in latent space.
• Transfer: from one site's memory to a model of all sites and cross-domain representations.
• Research direction: design experiments and decide what to ship.

Conocimientos

World models

Self-supervised learning

Probabilistic modelling

Educación

PhD in ML/CV/Robotics

Herramientas

PyTorch

Descripción del empleo

Laelaps AI Ltd. in Zurich is seeking a research scientist to lead work on world models, self-supervised learning, and probabilistic spatio-temporal modelling for our LEA foundation model of the physical world.

You will advance site memory, anomaly detection, and multi-sensor data integration in production-ready systems. You will collaborate with PhD-level founders, design experiments, and decide what ships.