Foundation Model AI Architect (Molecular & Multimodal Systems)
Our client is exploring a new generation of AI architectures grounded in principles from computational neuroscience, biological computation, and multimodal modelling. Their aim is to build large foundation models capable of reasoning over molecular, structural, and scientific datasets with explainability and precision.
They are hiring a Foundation Model AI Architect to lead the design of advanced neural systems that combine transformer architectures, causal reasoning models, multimodal representations, and agentic behaviours. You will design models that integrate chemical data, molecular structures, spectroscopic signatures, and simulation derived information into unified AI systems for materials discovery.
A key component of this role involves scaling models on high performance GPU clusters, optimising training and inference pipelines, and working with advanced frameworks such as JAX. You will also build automated labelling systems, behavioural encoding workflows, and interpretable ML pipelines that support transparency and scientific trustworthiness.
This position suits someone who can translate ideas from systems neuroscience and complex biological modelling into practical, engineered AI architectures for real scientific problems.
Ideal Profile:
PhD in computational neuroscience or computational biology, deep expertise in neural architecture design, strong GPU/HPC programming skills, and experience developing large scale or multimodal foundation models.
