Train, fine-tune, and evaluate large-language / foundation models; apply parameter-efficient techniques, reinforcement learning, and prompt engineering.
Build and enhance memory and context management systems, manage embeddings and retrieval pipelines.
Integrate structured knowledge sources like knowledge graphs and ontologies to improve reasoning and reduce hallucinations.
Develop pipelines for ingesting and processing unstructured and multi-format data types, including parsing, embedding, indexing, and retrieval.
Optimise inference and system performance to reduce latency and increase throughput; design efficient prompt pipelines.
Define and implement monitoring and evaluation frameworks tracking accuracy, drift, error rates, and user satisfaction.
Collaborate with engineers, product managers, and UX teams to convert experiments into scalable features.
Stay hands-on: write code, prototype solutions, and contribute directly to implementation (including some Java).
Requirements
Proven experience working with large language / foundation models, including training, tuning, and applying modern GenAI techniques such as reinforcement learning, memory and context systems, retrieval augmentation, structured knowledge integration, optimisation, and evaluation.
Master’s degree in a relevant field required; PhD preferred.
Strong coding skills; Java is highly desirable.
Track record of delivering software features or products in a collaborative environment, ideally within a product-led organisation.
Passion for emerging AI technologies and ability to translate research into practice.
Comfortable working cross-functionally to move from prototypes to production-ready features.