Machine Learning Researcher | Protein Design | Diffusion Modelling | Flow Matching | Zurich, Switzerland in Zürich

About the role

An innovative, high-growth AI-driven biotechnology company is looking for an experienced Machine Learning Researcher specialising in protein structure modelling to help develop advanced machine learning systems for protein discovery, engineering and optimisation.

The company is building a machine learning platform that combines computational modelling with experimental validation, supporting applications across industries including biopharma, agriculture, food and industrial biotechnology.

In this role, you will contribute to the structural modelling capabilities underpinning both protein engineering and de novo protein design. Key research areas include:

• Structure and complex prediction: Training and fine-tuning protein structure prediction and co-folding models.




• Structure-conditioned property prediction: Developing models that use predicted and experimental structures to predict multiple protein properties.




• Structure representation learning: Working with equivariant architectures and structural encoders.




• Active learning: Designing and implementing active learning strategies to optimise the design-build-test-learn cycle and maximise the value of experimental data.

You will be expected to push the boundaries of machine learning research in protein engineering and de novo design. Working closely with research scientists and engineers, you will take novel approaches from initial research through to robust, production-ready implementations, with a strong emphasis on scalability and software engineering standards.

Research environment

A key advantage is access to a dedicated in-house wet lab, enabling rapid experimental validation of protein designs and research hypotheses. Short experimental turnaround times and high-throughput capabilities allow researchers to test ideas quickly and generate valuable proprietary datasets.

This creates a direct feedback loop between computational research and experimental biology, allowing you to assess the real-world performance of your models and continuously improve them.

Requirements

Essential technical experience

• PhD in Computer Science, Mathematics, Physics, Structural Biology, Computational Biology or a related discipline.




• Hands‑on experience building and training protein structure models, rather than simply running existing tools. Relevant experience includes structure or complex prediction, co‑folding, backbone generation, inverse folding and structure‑conditioned property prediction.




• Strong knowledge of geometric deep learning for molecules, including equivariant architectures, frame‑or‑coordinate‑based representations, and attention mechanisms over 3D structures.




• Experience with modern generative modelling techniques, such as diffusion models, flow matching and masked language modelling.




• Strong statistical and applied mathematical skills, particularly in evaluating results from small, noisy or confounded datasets.




• Familiarity with PyTorch or JAX.

Essential personal attributes

• Self‑directed: Comfortable tackling abstract research problems independently, structuring your work and delivering results without constant supervision.




• Growth‑oriented: Motivated by challenging technical problems, continuous development and working in a fast‑moving environment.




• Strong communicator: Able to communicate effectively across disciplines, particularly at the intersection of biology and software engineering.




• Collaborative: A kind, supportive team player who contributes to a positive and constructive working environment.




• Large‑scale model training, including multi‑node and sharded training, and managing long‑running training workloads.




• ML infrastructure, model provisioning, orchestration and deployment tools, such as Kubernetes and MLflow.




• Programming languages beyond Python, particularly Rust, Haskell, Julia or Elixir, and experience with functional programming paradigms.




• Inference optimisation, including Triton, Pallas and custom attention kernels for structural data.




• Experience taking deep learning models from research into production.




• Experimental structural biology, including crystallography, cryo‑electron microscopy (cryo‑EM) or NMR, and understanding how experimental artefacts affect training data.




• Physics‑based modelling, including molecular dynamics, docking and force fields, with an understanding of their strengths and limitations relative to learned models.




• Familiarity with protein design and synthetic biology applications, such as bispecifics, CAR‑T and metabolic engineering.

The opportunity

This is an opportunity to work at the intersection of machine learning research, computational protein design and experimental biology , developing novel models with a direct route to experimental validation.

You will have the opportunity to shape research direction, work on technically challenging problems and help translate cutting‑edge machine learning methods into practical protein engineering applications.

#J-18808-Ljbffr