Data Scientist III

RELX Group; Poltimore; On-site; GBP 65,000 - 90,000; Full time

Benefits offered by this job

Generous holiday allowance

Private medical benefits

Pension scheme

Share option scheme

Travel season ticket loan

Electric Vehicle Scheme

Dental Insurance

Life assurance

Employee learning & development

Job summary

LexisNexis Intellectual Property Solutions (LNIP) seeks a Data Scientist III to design and validate AI-driven capabilities for Protégé in PatentSight. You will lead NLP/LLM modeling, RAG prompts, and hybrid search to tackle multi-step IP tasks and deliver actionable insights for patent workflows.

You will partner with engineers to translate experiments into scalable, customer-facing solutions, establishing rigorous evaluation and benchmarking to influence product strategy and business impact.

Qualifications

• Degree in a quantitative or technical field (Statistics, Computer Science, Mathematics, Data Science).
• Strong experience in machine learning, NLP, and LLM-based modeling.
• Strong experience designing and running experiments, including model evaluation and coding skills in Python.
• Experience with generative AI techniques (e.g., prompt engineering, RAG).
• Experience designing and evaluating hybrid search (semantic + lexical) using embeddings and vector designing agent workflows, with hands-on experience applying agent frameworks.
• Proficiency in data analysis tools.
• Strong foundation in statistics, modeling, and large-scale text processing.

Responsibilities

• Develop and implement NLP, LLM, and generative AI approaches (e.g., RAG, prompt strategies, patent search).
• Define agentic workflows and reasoning strategies for multi-step IP tasks.
• Develop retrieval strategies, including hybrid search (semantic + lexical), and evaluation metrics (e.g., relevance, ranking quality).
• Analyse large-scale IP datasets to extract insights and improve model performance.
• Establish best practices for model evaluation, validation, and benchmarking.
• Translate experimental results into clear product recommendations and business impact.

Skills

NLP

LLM modeling

Machine learning

Python

RAG

Prompt strategies

Patent search

Experiment design

Embeddings

Hybrid search

Education

Degree in a quantitative field

Statistics

Computer Science

Data Science

Tools

LangChain

LangGraph

AutoGen

Google ADK

Embeddings

Job description

a Data Scientist III to help the design and validation of Agentic AI-driven product capabilities within Protégé in PatentSight. This role focuses on defining what to build and why, leveraging machine learning, NLP, and large language models (LLMs) to solve complex Intellectual Property (IP) workflows. You will help drive experimentation, modeling, and evaluation, partnering closely with engineers to translate validated approaches into scalable, customer-facing solutions.Key Responsibilities:Develop and implement NLP, LLM, and generative AI approaches (e.g., RAG, prompt strategies, patent search).Define agentic workflows and reasoning strategies for multi-step IP tasks.Develop retrieval strategies, including hybrid search (semantic + lexical), and evaluation metrics (e.g., relevance, ranking quality).Analyse large-scale IP datasets to extract insights and improve model performance.Establish best practices for model evaluation, validation, and benchmarking.Translate experimental results
into clear product recommendations and business impact.

Key Responsibilities:

• Develop and implement NLP, LLM, and generative AI approaches (e.g., RAG, prompt strategies, patent search).
• Define agentic workflows and reasoning strategies for multi-step IP tasks.
• Develop retrieval strategies, including hybrid search (semantic + lexical), and evaluation metrics (e.g., relevance, ranking quality).
• Analyse large-scale IP datasets to extract insights and improve model performance.
• Establish best practices for model evaluation, validation, and benchmarking.
• Translate experimental results
  into clear product recommendations and business impact.

Requirements:

• Degree in a quantitative or technical field (Statistics, Computer Science, Mathematics, Data Science, etc.).
• Strong experience in machine learning, NLP, and LLM-based modeling.
• Strong experience designing and running experiments, including model evaluation and coding skills in Python.
• Experience with generative AI techniques (e.g., prompt engineering, RAG).
• Experience designing and evaluating hybrid search (semantic + lexical) using embeddings and vector designing agentic workflows and reasoning strategies, with hands-on experience applying agent frameworks (e.g., Google ADK, LangChain, LangGraph, AutoGen) in real-world use cases.
• Proficiency in data analysis tools.
• Strong foundation in statistics, modeling, and large-scale text processing.

Why Join Us?

Join our team and contribute to a culture of innovation, collaboration, and excellence. If
Data Scientist IIIExcited to uncover insights from complex data and turn them into impactful business decisions?Do you enjoy building predictive models, experimenting with advanced analytics, and solving real-world challenges with data?

About the team:

LexisNexis Intellectual Property Solutions (LNIP) is the global leader in patent intelligence, bringing clarity to innovation for businesses, law firms, universities, and government agencies worldwide. Our mission is to help the innovation community make better decisions faster, with greater confidence, combining the world's most trusted patent data with sophisticated analytics, AI-powered insights, and purpose-built workflows.Protégé in PatentSight is LNIP's next-generation agentic AI assistant, purpose-built for strategic patent analysis. Protégé replaces complex filter-based workflows with natural language questions, surfacing structured, decision-ready insights grounded in trusted data and established metrics.

About the role:

We are seeking you are ready to advance your career and make a significant impact, we encourage you to apply.

Work in a way that works for you

We promote a healthy work/life balance across the organization. We offer an appealing working prospect for our people. With numerous wellbeing initiatives, shared parental leave, study assistance and sabbaticals, we will help you meet your immediate responsibilities and your long-term goals. Working flexible hours - flexing the times when you work in the day to help you fit everything in and work when you are the most productive.Working for youWe know that your well‑being and happiness are key to a long and successful career. These are some of the benefits we are delighted to offer:Generous holiday allowance with the option to buy additional days.Health screening, eye care vouchers, and private medical benefitsWellbeing programsLife assuranceAccess to a competitive contributory pension schemeSave As You Earn share option scheme.Travel Season ticket
loan.Electric Vehicle SchemeOptional Dental InsuranceMaternity, paternity, and shared parental leaveEmployee Assistance ProgrammeAccess to emergency care for both the elderly and childrenRECARES days, giving you time to support the charities and causes that matter to you.Access to employee resource groups with dedicated time to volunteer.Access to extensive learning and development resourcesAccess to employee discounts scheme via Perks at WorkAbout the business-At LexisNexis Intellectual Property (LNIP), we believe that whenever a person works on a patent and understands the future trajectory of a specific technology, that person has the potential to fundamentally change how society operates. We are proud to directly support and serve these innovators in their endeavors to better humankind. We enable innovators to accomplish more by helping them make informed decisions, be more productive, comply with regulations, and ultimately achieve superior results. By harnessing the latest advances in machine learning combined with expert analysis, LexisNexis Intellectual Property is disrupting how actionable insight is extracted from patent data. Information can now be accessed with efficiency, accuracy and at a speed that is just not possible by traditional methods. Our overall success is measured by how well we deliver these results.