Role Summary
We are building Sibylline’s core data and intelligence platform to support real-world security, geopolitical, and risk decision-making. We are seeking a Senior Data Scientist specialising in applied modelling to develop analytical models and AI-enabled methodologies that combine quantitative data, qualitative assessment, and expert judgement.
This role focuses on building indices, indicators, trend analyses, and structured assessments across areas such as geopolitics, crime, security, public health, and societal risk. A key part of the role involves using Large Language Models (LLMs) to support qualitative analysis at scale, to, for example, extract signals from text, structure expert assessments, enrich data, and support transparent, auditable analytical workflows.
Your work will be embedded directly into a developing enterprise SaaS intelligence platform to be used by analysts and clients globally.
This is a hands-on role suited to someone who enjoys working at the intersection of data science, qualitative analysis, and applied AI, and who is comfortable operating in complex, ambiguous real-world domains.
This is a remote role working 9 am - 5 pm, Monday to Friday, requiring occasional travel to London. The London office is available to work in if desired.
Responsibilities
- Designing and maintaining quantitative models, indices, and scoring frameworks for diverse areas such as geopolitical, security, crime, and societal risk
- Developing methodologies that combine structured qualitative judgement with quantitative data
- Using LLMs to support qualitative analysis at scale (e.g. text analysis, signal extraction, summarisation, classification, enrichment)
- Designing AI-enabled analytical workflows that are transparent, explainable, and suitable for client-facing intelligence products for internal and external products
- Analysing trends across time and geography using heterogeneous data sources (structured, semi-structured, and unstructured)
- Validating, stress-testing, and iterating models as new data and intelligence become available
- Clearly documenting assumptions, methodologies, and limitations for both technical and non-technical stakeholders
- Working closely with intelligence analysts, product managers, and software engineers to embed models and AI workflows into production systems
- Contributing to the development of Sibylline’s analytical, modelling, and AI governance standards
Knowledge, Skills, and Abilities
- Strong experience in applied data science, quantitative modelling, or analytical research in real-world decision-making contexts
- Desired: prior experience in geopolitical or geoeconomic modelling
- Proficiency in Python for data analysis and modelling
- Working knowledge of SQL and similar technologies for querying and manipulating data
- Desirable: knowledge of frontend development including React, Typescript
- Experience building indices, indicators, forecasts, or composite scoring systems
- Ability to structure and formalise qualitative judgement and expert assessment
- Hands-on experience using LLMs in applied analytical workflows (e.g. prompt design, evaluation, chaining, or hybrid human–AI analysis)
- Strong statistical reasoning and data analysis skills
- Proficiency in Python (or similar) for data analysis and modelling
- Experience working with time-series, geospatial, event-based, or text data is highly desirable
- Prior experience in geopolitical or geoeconomic modelling is highly desirable
- Strong interest in geopolitics, security, crime, public health, or societal risk
- Comfortable working in a trusted environment and eligible for background screening
Qualifications
- Degree in a quantitative or analytical discipline (e.g. data science, statistics, economics, mathematics, political science with quantitative focus, epidemiology, or similar)
- Experience working with intelligence, risk, or security-related datasets
- Familiarity with model interpretability, explainability, and AI governance in applied settings
- Experience collaborating with software engineering teams to productionise analytical and AI-driven outputs