NLP Data Scientist, Quantexa

• Degree in computer science, computational linguistics, artificial intelligence, mathematics, physics, engineering, or related field
• Familiarity with NLP and machine learning proven through personal, university or work projects
• Experience building or adapting machine learning models end-to-end, with the ability to explain the decisions made
• Proficiency in Python (or similarly complex programming language), which is the primary programming language used
• Willingness to learn about unfamiliar languages and scripts

Desirable

• We're hiring an NLP Data Scientist to work on challenging NLP problems that underpin Quantexa’s entity resolution product: messy and multilingual text, creating training data from scratch, and careful evaluation
• The role sits in our NLP Centre of Excellence, which builds NLP capabilities for Quantexa's products - an increasingly important area as the platform is applied to unstructured text
• Our work ranges from shipping production models for tasks like named entity recognition, relation extraction, and text classification, through to more exploratory research that drives improvements in our products and the wider field. We work on the full spectrum of model complexity - from fast, explainable models, to fine-tuned small language models to agentic LLM-based systems
• As a NLP Data Scientist, you'll develop these NLP components end-to-end from framing the problem and building training and testing datasets, training and evaluating models, handover for production, and improving accuracy based on feedback
• Near-term, this is focused on building lightweight, fast NLP models where careful design of training data and evaluation benchmarks is critical
• Translating underspecified business problems into well-scoped ML tasks
• Training, evaluating, improving NLP models end-to-end through reproducible pipelines
• Sourcing, building, curating training and evaluation datasets from heterogeneous, often imperfect inputs
• Working across languages and scripts, such as Arabic, Japanese, Chinese, and others
• Developing and using LLM-assisted workflows for dataset generation, error analysis, and model improvement iterations
• Collaborating with engineering and product teams to turn ideas into integrated product features; presenting findings to stakeholders