Data Scientist

Data Scientist
You are a machine learning expert with experience building real-world systems under technical, operational, and security constraints. You're skilled in building NLP workflows and are comfortable experimenting with LLMs in bespoke settings. Your
solutions must be robust, auditable, and interoperable within complex human systems.

This role focuses on delivering fit-for-purpose ML solutions for experimental AI applications in Defence. You'll be working on tasks such as transcription analysis, topic discovery, graph-based reasoning, and language model alignment. These solutions
must function at the edge, in secure and air-gapped environments. You'll work closely with a cross-functional team of backend engineers, frontend engineers, platform engineers and external stakeholders.

What You'll Do
* Design and implement NLP pipelines for multi-speaker audio transcription, topic extraction, summarisation, and insight generation
* Develop modular LLM workflows (e.g., RAG, graph-based reasoning) tailored for resource-constrained deployments
* Build and test solutions that incorporate structured and unstructured data, such as event transcripts, policy documents, and supply chain graphs
* Work with infrastructure teams to containerise and optimise ML models for edge inference
* Help define robust evaluation protocols using both quantitative metrics and human feedback
* Collaborate on secure deployment approaches that respect classification boundaries and deployment restrictions
* Provide technical insights during experimental exercises, including on-site troubleshooting and performance tuning
* Support the delivery team in aligning AI outputs with the qualitative and analytical needs of decision-making activities

What You Bring
* Practical experience training and deploying NLP models, including LLMs, summarisation, entity extraction, and topic modelling
* Strong Python skills with proficiency in ML tooling (e.g., HuggingFace Transformers, spaCy, PyTorch/TensorFlow, RAG frameworks)
* Experience working with knowledge graphs, graph neural networks, or graphenabled retrieval workflows
* Demonstrated ability to build ML solutions that can function in edge, lowcompute, or air-gapped environments
* Familiarity with evaluation methodologies that include human-in-the-loop and subject matter expert review
* Bonus: Experience working in Defence, aerospace, or other secure environments


Due to the nature and urgency of this post, candidates holding or who have held high level security clearance in the past are most welcome to apply. Please note successful applicants will be required to be security cleared prior to appointment which can take up to a minimum 10 weeks.