You’ll join our Core Foundation Model Team—a transversal group developing the architectures, training methodologies, and abstractions powering all our domain-specific squads. This team works at the heart of our model ecosystem, enabling scalable, reusable research across verticals.

We’re still early. You’ll be among the first hires (joining Mike Gartrell), helping define the trajectory of our generative modeling capabilities from the ground up.

As a Generative Time Series Researcher, you’ll focus on modeling and generating structured temporal data across a wide range of scientific and real-world domains—ranging from brain signals (EEG, MEG, fMRI) to complex data streams in in various scientific and industrial areas.

This is a cross-functional, deep research role at the intersection of machine learning, temporal modeling, and domain-specific science. You’ll blend theoretical innovation with real-world relevance and play a key role in unlocking simulation, reconstruction, and augmentation capabilities across our stack.

• Design, prototype, and benchmark generative models for irregular, noisy, and multimodal time series (e.g., diffusion models, latent ODEs, temporal VAEs).

• Advance architectures for long-context and high-resolution modeling, including transformer-based and memory-augmented methods.

• Build abstractions and tooling for training large-scale generative models on scientific datasets (e.g., tokenization, augmentation, masking strategies).

• Create domain-agnostic frameworks for time series modeling that can generalize across use cases (Brain, Spine, Industrial, etc.).

• Explore intersections with causality, simulation, and uncertainty quantification to build models that support scientific and operational inference.

• Contribute to scientific publications (e.g., NeurIPS, ICLR, ICML, AISTATS) and help grow our open research presence.

• Collaborate with domain-focused squads to integrate generative capabilities into applied pipelines.

• PhD in Machine Learning

• Solid track record in generative modelling, especially for time series or dynamical systems.

• Fluency in probabilistic modelling, diffusion models, transformers, or neural ODEs.

• Experience working with real-world time series data (e.g., biomedical signals, finance, physics).

• Proficiency in Python and ML frameworks (e.g., PyTorch, JAX).

• Strong publication record and ability to translate research into usable code and systems.

• Bonus: experience with multimodal learning (e.g., EEG + fMRI), foundation model training, or infrastructure at scale.

• Application review (mid-August, after Paul’s holiday break)

• Introductory call with Paul (Head of Talent Acquisition) – 30 min

• Deep dive on AI research with Mike Gartrell – 45 min

• Behavioural interview with Paul – 45 min

• Research talk, coding/pair design session, and team discussions – Half-day onsite or remote