Machine Learning Researcher
About the Role
We are a rapidly growing technology company developing advanced machine learning systems for biological research and protein engineering.
We are looking for a Machine Learning Researcher with substantial experience in protein structure modelling to contribute to the development of ML systems for iterative, experiment-driven protein discovery and optimisation.
Our technology is used across a range of life-sciences and industrial applications, supporting organisations working on complex protein engineering and design challenges.
This role will contribute primarily to the structural side of our modelling stack, spanning both protein engineering and de novo protein design.
Key areas of research include:
- Structure and complex prediction: Training, fine-tuning and developing models for protein structure prediction, complex prediction and co-folding.
- Generative structure design: Developing methods for de novo backbone generation, inverse folding, motif scaffolding and molecular interaction design.
- Structure-conditioned property prediction: Using predicted and experimental structural information within models that predict multiple relevant protein properties.
- Structure representation learning: Developing geometric and equivariant architectures and representations for three-dimensional biological structures.
- Active learning: Designing iterative learning strategies that incorporate experimental results into subsequent modelling and design rounds, maximising the information gained from each round of data.
You will be expected to advance the state of the art by researching and developing novel machine-learning approaches to protein engineering and de novo design.
You will work closely with other researchers, software engineers and scientists to take ideas from early-stage research through to robust, scalable implementations suitable for production environments.
The company has access to dedicated experimental capabilities, enabling computational designs and research hypotheses to be tested rapidly. This creates a close feedback loop between machine-learning research and experimental results and provides access to proprietary datasets generated through iterative design and validation.
Requirements
Essential Technical Experience
- PhD in Computer Science, Mathematics, Physics, Structural Biology, Computational Biology, or another relevant quantitative discipline.
- Excellent software development skills.
- Hands-on experience building and training protein structure models, rather than solely using existing models or tools.
- Relevant experience could include structure or complex prediction, co-folding, backbone generation, inverse folding, molecular interaction design, or structure-conditioned property prediction.
- Experience with geometric deep learning for molecular or biological systems, including areas such as equivariant architectures, coordinate- or frame-based representations, and attention mechanisms operating over 3D structures.
- Knowledge of modern generative modelling approaches, such as diffusion models, flow-based methods or masked modelling.
- Strong statistics and applied mathematics skills, particularly for evaluating results using small, noisy or potentially confounded experimental datasets.
- Familiarity with modern machine-learning frameworks such as PyTorch or JAX.
Essential Non-Technical Skills
- Self-directed: Comfortable tackling open-ended research problems, organising your own work and delivering results with a high degree of autonomy.
- Growth-oriented: Motivated by challenging technical problems and comfortable receiving and acting on constructive feedback.
- Strong communicator: Able to communicate effectively across machine learning, software engineering and biological science disciplines.
- Collaborative: Works effectively within interdisciplinary teams and contributes positively to the wider working environment.
Nice to Have
- Experience with large-scale or distributed model training, including multi-node or sharded training and managing long-running experiments.
- Familiarity with machine-learning infrastructure, model pipelines and deployment technologies such as Kubernetes, ML experiment-management platforms or equivalent tooling.
- Experience with programming languages beyond Python, particularly languages or software architectures influenced by functional programming.
- Experience optimising inference performance, including specialised GPU programming, custom kernels or optimisation of attention mechanisms for structural models.
- Demonstrated experience deploying deep-learning models into production environments.
- Knowledge of experimental structural biology techniques such as crystallography, cryo-EM or NMR, including an understanding of how experimental artefacts can affect machine-learning datasets.
- Experience with physics-based computational modelling, such as molecular dynamics, molecular docking or force-field methods, and an understanding of where these approaches complement learned models.
- Familiarity with practical protein engineering, therapeutic development, synthetic biology or related biotechnology applications.

Why Join?
- Work on technically challenging problems at the intersection of machine learning, biology and scientific discovery.
- Develop novel models that can be tested against real experimental data.
- Work within a highly interdisciplinary team spanning machine learning, engineering and biological science.
- Help shape a growing technology platform and research programme.
- Take research ideas from initial exploration through experimental validation and ultimately into production use.