About the role
Responsibilities:
- Design and implement continual learning algorithms for production systems.
- Develop sample-efficient training methods that can learn from sparse interaction signals.
- Build retrieval and memory systems that adapt in real time.
- Conduct experiments, analyse results and iterate quickly on research ideas.
- Optimise machine learning pipelines for latency and cost at scale.
- Contribute to technical direction and architecture decisions.
Requirements:
- Strong fundamentals in Machine Learning, including optimisation, gradient methods and neural architectures.
- 2+ years of hands-on experience training and deploying machine learning models.
- Strong proficiency in Python and experience with PyTorch, JAX or an equivalent framework.
- Experience with retrieval-augmented systems or reinforcement learning.
- Ability to read research papers, implement proposed ideas and conduct rigorous experiments.
- Publications at leading ML research conferences such as NeurIPS, ICML, ICLR, ACL or similar venues.
Nice to Have:
- Experience with large-scale distributed training.
- Experience with personalisation, recommendation systems or ranking.
- Familiarity with LoRA, adapters or parameter-efficient fine-tuning techniques.
- Experience building or working with continual learning systems.
- Experience deploying and maintaining production ML systems serving real users.