Remote • London
Designingarchitectingand implementing large-scale foundation modelsIn-depth knowledge of recent architectures (e.g.MoEstate-space models)Understanding of data curation and quality control for massive training datasetsProficiency in CUDA C/C++ programmingProficiency in PythonOptimising and debugging deep learning models and performance bottlenecksExperience with distributed or large-scale training and inferenceDeep understanding of at least one major deep learning framework (ideally PyTorch)Experience with cloud platforms (e.g. AWSAzureGCP)ContainerisationOrchestrationExperiment trackingMonitoringEvaluation pipelinesPassion and determinationProblem-solving skillsDelivery-orientedOpenness to disagreement