Applied AI Scientist
You will research, design, and develop deep learning models using transformer architectures, PyTorch, TensorFlow, and Python, plus AWS SageMaker and S3. The team builds foundation models from large-scale transactional and behavioural datasets to power customer intelligence for a major UK bank.
Free Tailor for ATS: 10/10 runs left
Applied AI Research Scientist – London or Edinburgh (Remote)
Competitive Six Figure Package
AI Connect are partnering with a major UK bank to help them appoint an Applied AI Research Scientist to join a growing Applied AI function developing next-generation AI capability.
You'll join a highly technical research team responsible for designing, developing and evaluating modern deep learning models that underpin customer intelligence across the organisation. Working with large-scale transactional and behavioural datasets, you'll help develop transformer-based foundation models that support a wide range of downstream AI applications.
This is a genuine applied research role, combining scientific rigour with real-world business impact.
What You'll Do
Research, design and develop modern deep learning models using transformer-based architectures.
Build and evaluate Foundation Models that learn rich customer representations from transactional data.
Design, implement and optimise training pipelines using frameworks such as PyTorch and TensorFlow.
Experiment with different model architectures, training objectives and evaluation techniques to improve model performance.
Work with embeddings, representation learning and transformer methodologies to solve complex AI problems.
Collaborate closely with Applied AI Engineers and ML Platform Engineers to transition research into production environments.
Contribute to technical direction through scientific experimentation, model evaluation and architectural decision making.
Key Skills & Experience
Strong understanding of modern Deep Learning methodologies and Transformer architectures.
Able to explain concepts such as tokenisation, embeddings, positional encoding, self-attention, multi-head attention and transformer training approaches.
Strong Python programming skills with hands-on experience using PyTorch, TensorFlow or similar deep learning frameworks.
Experience implementing, reviewing and extending deep learning codebases rather than simply consuming pre-trained models.
Ability to explain implementation decisions within transformer training pipelines and the trade-offs between different modelling approaches.
Understanding of cloud-based machine learning environments, including AWS services such as SageMaker, S3 and modern MLOps concepts.
Experience evaluating AI models using robust scientific methodology and experimentation.