ML/AI Research Scientist

Engitix Therapeutics is a London-based biotechnology company unlocking the extracellular matrix (ECM) as a source of novel therapeutic targets for fibrosis and cancer. Our proprietary platform integrates proteomics, and transcriptomics data on patient tissue samples with rich clinical metadata to identify and validate targets in the tumour microenvironment and fibrotic tissue.

We are working to build a first-of-its-kind AI/ML capability to transform how we extract biological insight from complex proteomic data, with a focus on developing foundation models for mass spectrometry that go beyond existing analytical pipelines.

The opportunity:
This will be a founding role in Engitix’s AI/ML research programme. You will take a leading role in the design and development of foundational models for proteomics analysis. This will involve developing novel architectures seeking to map and make use of approximately 60–70% of current proteomics data that remains unexplained. Our existing explorations have focused on self-supervision, and object centric learning e.g. slot-attention, though this by no means is set in stone. Success will build representations that dramatically improve peptide identification, quantification, and discovery of novel biology. There will also be ample opportunity to work on problems outside the proteomics domain if interested.

Your responsibilities:

• Lead the research, design, and implementation of foundational models for mass spectrometry data analysis, focused on proteomics
• Test and optimize performance on small and large-scale training datasets from public spectral repositories and internal Engitix data
• Benchmark against state-of-the-art tools (DIA-NN, Spectronaut, MSFragger-DIA, MaxDIA)
• Design active learning and experimental design strategies that close the loop between model predictions and wet-lab validation
• Publish at top-tier venues (NeurIPS, ICML, ICLR) and contribute to the open scientific community
• Shape the long-term AI/ML research roadmap at Engitix

You might be finishing a PhD or postdoc at a top ML or computational biology group. You might be 2–7 years into an industry research role and looking for something more impactful and autonomous. You might be a generative modelling expert who’s never touched biology but is excited by the idea of building a foundation model for a new data modality. Or you might be a computational biologist who’s been publishing at NeurIPS and wants to apply your skills to a real drug discovery programme.

What matters most is that you are an excellent ML researcher with a track record of rigorous, published work, and that you are genuinely motivated by the opportunity to build something new at the intersection of deep learning and biology.

Required:

• PhD in machine learning, computer science, computational biology, statistics, physics, or a related quantitative field
• Strong publication record at top-tier ML conferences (NeurIPS, ICML, ICLR) and/or leading scientific journals (Nature, Nature Methods, Nature Biotechnology, Nature Machine Intelligence)
• Deep expertise in at least one of: self-supervised learning, transformer architectures, attention mechanisms, generative models (diffusion, flow matching, VAEs), representation learning, object-centric learning
• Strong implementation skills in PyTorch (or JAX); experience training models on GPUs at scale
• Genuine intellectual curiosity about biological data and a desire to work at the interface of ML, biology, and therapeutics discovery

Nice to have:

• Experience with mass spectrometry data (proteomics, metabolomics, or small-molecule MS/MS)
• Familiarity with computational proteomics pipelines (DIA-NN, Prosit, Spectronaut, Percolator, or similar)
• Experience building foundation models or large-scale self-supervised pretraining systems
• Background in spectral data, signal processing, or time-series modelling
• Understanding of protein biology, sequence models (ESM, MSA Transformer), or structural biology
• Experience with multi-modal or cross-modal learning (e.g., contrastive learning across modalities)
• Track record of bridging ML research with real-world biological or clinical applications
• Exposure to drug discovery, single-cell or spatial transcriptomics data