Skip to content

    AI for life sciences, built by a bioinformatician.

    The same four engagements, applied to biotech, pharma, and healthcare problems we've worked inside.

    From single-cell and spatial omics to multi-omics integration and LLM-driven literature work — we scope, plan, or build the parts where AI is actually worth the effort, and tell you where it isn't.

    Talk to us about a life-sciences project

    Heavy data, real biology

    Omics scale, complex assay outputs, and clinical metadata that doesn't merge cleanly. Standard analysis stops being enough fast.

    Evidence, not enthusiasm

    Every recommendation starts with a data-readiness check. Most AI ideas die or pivot here — and that saves you the bigger spend later.

    PhD-grounded

    Bioinformatics, machine learning, systems biology. The founder's actual research domain, not a vertical we added to the pitch deck.

    What domain expertise changes in practice

    A generalist data scientist will normalise single-cell RNA-seq the way they'd normalise bulk RNA-seq. It looks fine for two weeks, then the clustering quietly misleads you. A generalist will treat a batch effect as noise to remove, where a biologist knows it might be the signal. The difference between a domain specialist and a capable generalist rarely shows up in the demo — it shows up six months later, in whether the result held.

    Where life-science teams typically engage us

    Pick the entry point that matches your situation. Each maps to one of the four engagements.

    Single-cell & spatial omics

    End-to-end pipelines for scRNA-seq, scATAC-seq, CITE-seq, Visium, MERFISH. QC, clustering, annotation, neighbourhood analysis, cell-cell interaction modelling — readable by both your wet-lab team and your CSO.

    Typical engagement: Scoping Lab or Pilot Build

    Multi-omics integration

    Harmonising bulk, single-cell, proteomic and clinical data into queryable models. Network analysis, mechanistic modelling, biomarker discovery — built so a translational team can act on the output.

    Typical engagement: Scoping Lab or Pilot Build

    Generative AI for R&D

    LLM-driven literature review, hypothesis generation, protocol drafting, and agent-based experiment orchestration. Built with traceable prompts, validated outputs, and human-in-the-loop checkpoints.

    Typical engagement: Workshop or Scoping Lab

    AI strategy for regulated life-sciences

    Roadmaps and use-case prioritisation for biotech, pharma and healthcare teams operating under the EU AI Act, EMA guidance, GDPR for health data, and MDR where applicable. EU data residency by default.

    Typical engagement: Workshop or Roadmap

    What we work with

    Concrete tools and modalities — not a list of buzzwords.

    Single-cell & spatial

    • Scanpy
    • Seurat
    • Squidpy
    • AnnData

    Pipelines & infra

    • Nextflow
    • Snakemake
    • Bioconductor
    • Docker

    ML & generative

    • PyTorch
    • JAX
    • scikit-learn
    • HuggingFace
    • MONAI

    Modalities

    • scRNA-seq
    • scATAC-seq
    • CITE-seq
    • Visium
    • MERFISH
    • bulk RNA-seq

    A short call beats a long brief.

    Tell us where you are — we'll tell you which engagement (if any) matches, and what a useful first step looks like.