Senior Director of Causal & Mechanistic AI — a pivotal scientific leadership role shaping how next-generation causal discovery is applied to human bio
Senior Director, Causal & Mechanistic AI
Location: London (Hybrid)
Search Partner: KEMIO Consulting (Retained)
Company: Stealth TechBio Start-up
KEMIO Consulting has been exclusively retained by a stealth-stage TechBio company in London to appoint a Senior Director of Causal & Mechanistic AI — a pivotal scientific leadership role shaping how next-generation causal discovery is applied to human biology and genomics for therapeutic insight.
This is not a “feature engineering” or production ML role — it is a first-principles scientific research mandate to advance the frontier of causal inference in messy, high-dimensional biological systems and translate those discoveries into target and biomarker hypotheses that genuinely hold up under experimental validation.
You will build a team that develops new methods that combine causal discovery + mechanistic reasoning to uncover disease-driving biology with quantified uncertainty — enabling target identification, biomarker discovery, and patient stratification that is both scientifically defensible and experimentally grounded.
We are looking for a senior researcher with deep expertise in modern machine learning, ideally across advanced deep generative modelling, variational inference, graph-based methods or neural causal models, with hands-on implementation experience in PyTorch or JAX and a track record of publication at top-tier venues such as NeurIPS, ICML or UAI.
You should have a strong grounding in causal inference, including a rigorous understanding of identifiability constraints, when causal structure can genuinely be learned, and how to quantify uncertainty in a principled way. The ideal candidate should also bring fluency in real biological data, particularly CRISPR screens or single-cell genomics, and understands the “messiness” of experimental evidence ,including confounding, batch effects and incomplete perturbations, together with experience collaborating on or informing experimental validation.
Experience in the areas below would be highly advantageous: