About the team

We are a research group at Aithyra in Vienna, Austria, at the intersection of machine learning and life sciences.

Our research focuses on generative modeling, flow models, optimal transport, and protein design. We develop novel machine learning algorithms with applications to understanding cellular dynamics, predicting protein structures, and designing new molecules.

Led by Alexander Tong, our group combines theoretical insights with practical applications to solve problems in computational biology and chemistry.

Research Focus and Collaborators

To see our main research directions check out: https://tonggroup.org/research/.

Within Aithyra, we currently have active collaborations with the Bronstein, Winter, Ceylan, and Mora groups. We collaborate internationally both on the methods and experimental sides.

Project 1: Controllable Generative Modeling

This PhD project will develop machine-learning methods for controllable generative modelling: the generation of new data subject to specified properties, constraints, or desired outcomes. The research will investigate how generative models can be guided more reliably and efficiently using conditioning information, structural constraints, interventions, rewards, or user-defined objectives.

Possible methodological directions include conditional diffusion and flow-based models, controllable sampling, representation learning, latent-space manipulation, constrained generation, reinforcement learning, guidance methods, and uncertainty-aware generation. The project will address fundamental questions concerning how control signals should be represented, how multiple objectives can be balanced, how generated outputs can be made interpretable and reliable, and how models can generalize to conditions not observed during training.

The project is not restricted to a single application domain. Potential use cases include molecular and protein design, scientific modelling, healthcare, robotics, language, vision, and other settings involving structured or sequential data. Applications will provide challenging test cases, while the primary goal will be to develop general and principled methods for controllable generation. To see more of what we are working on checkout our website: https://tonggroup.org/

Candidate’s Profile and Skills 1

What we’re looking for
› Strong background in machine learning, mathematics, or a related quantitative field
› Interest in applying ML to the life sciences — prior biology experience is welcome but not required
› Alignment with one or more of our research areas (generative modeling, flow models, optimal transport, protein design, single-cell biology)

Project 2: Generative Modelling for Molecular Design

This PhD project will develop and apply generative machine-learning models to design novel molecules and biomolecules with desired properties. The central aim is to enable computational models to generate candidate molecules while controlling characteristics such as structure, stability, activity, selectivity, solubility, or interactions with biological targets.

The project will combine generative AI with molecular and biological knowledge. Depending on the candidate’s interests and background, research may focus on protein or peptide design, small-molecule generation, molecular interactions, or the modelling of biological states and responses. The student will explore how experimental data, chemical principles, structural information, and functional measurements can be incorporated into generative models and used to guide the generation process. Generated candidates may be assessed through computational analyses, collaboration with experimental researchers, or laboratory validation.

The project is intentionally interdisciplinary and can be approached from either a computational or an experimental perspective. Its goal is to establish a more directed and useful approach to AI-assisted molecular discovery, in which models generate candidates that are not only novel but also relevant to a defined biological or chemical objective. To see more of what we are working on checkout our website: https://tonggroup.org/

Candidate’s Profile and Skills 2

No direct laboratory experience is required, although an understanding of molecular and biological concepts is important.