Machine Learning Scientist- Generative Modelling

Machine Learning Scientist, Generative Modelling

Location: London | Type: Full-time

About the company

A fast-growing TechBio company that develops new medicines by combining single-cell multi-omics from patient tissue, functional assays, and machine learning. The team is highly interdisciplinary, with biology, computation, and engineering working side by side in state-of-the-art wet and dry labs.

The opportunity

Help build the next generation of generative and predictive models of cellular behaviour. Your work will be central to understanding and controlling cellular decision-making, which enables new therapeutic strategies grounded in generative models. You’ll have access to cutting-edge multiomic and interventional datasets, advanced compute infrastructure, and modern ML tooling, including agentic workflows. You’ll also see your models tested directly in experimental biology.

What you’ll do

• Design and implement generative modelling approaches that learn intervention effects from diverse biological data, including single-cell perturbation experiments.
• Develop models that go beyond correlation, with a focus on generalisation, counterfactual prediction, and experimental design.
• Work with experimental teams to design and validate computational hypotheses, iterating to identify the highest-signal next experiment.
• Evaluate models for causal coherence, mechanistic fidelity, and real-world utility, not just fit.
• Communicate findings clearly to colleagues across disciplines.

What you’ll bring

• A PhD in ML, statistics, computer science, or a related quantitative field.
• Deep expertise in generative modelling.
• Strong foundations in probabilistic modelling, representation learning, or neural network architectures for structured or sequential data.
• Excellent Python skills and familiarity with scalable ML tooling and high-performance computing.
• A disciplined approach to model evaluation, including experiments that test real-world utility beyond standard benchmarks.
• Willingness to engage deeply with biological data. Experience with single-cell or perturbation datasets is a strong plus.

Nice to have

• A track record of impactful ML publications or open-source contributions.
• Experience in interdisciplinary teams or applying ML in real-world settings.