Location
London

Hours
Full Time

Salary
Competitive, commensurate with experience

About the Role
Theia Insights builds foundational financial intelligence products, including industry classification, knowledge graphs, and factor risk models, serving some of the largest asset managers, hedge funds, index providers, and sell-side banks. As a Quant Engineer on the Data Products team, you will build and maintain the models behind our Thematic Factor Risk Models (TFM), which decompose stock returns into thematic and traditional risk factors. Your work will include back-testing methodologies and turning research into daily production output in close collaboration with the economics team.
The Data Products team owns the data that underpins all our offerings. It is a small, senior group that prioritizes correctness and reproducibility over volume and works closely with product leads who shape the methodology.

Key responsibilities include:
- Developing statistical models of stock price movements and estimating thematic trend performance
- Constructing and back-testing factor risk models
- Designing and validating signal-generation and portfolio-attribution methodologies
- Ensuring research reproducibility so outputs can be re-run exactly, including after backfills and restatements
- Collaborating with the pipelines team to deploy modelling decisions into daily production

Experience
- Strong production Python skills
- In-depth knowledge of factor risk models and portfolio attribution including cross-sectional regression, covariance estimation and shrinkage, and rigorous back-testing
- Expertise in point-in-time discipline, avoiding look-ahead and survivorship bias, and reconstructing what was knowable on a given date
- Statistical modelling and optimization experience (e.g., statsmodels, cvxpy; PyTorch is a plus)
- Proficiency with datasets in pandas and Parquet/Arrow, and analytical engines such as DuckDB

About you
- Quantitative research background, either academic or industry, is desirable
- Familiarity with index construction and classification taxonomies
- Understanding of working with model-derived inputs, including NLP-generated exposures that carry measurement error and revise over time
- Experience with task orchestration tools (Dagster or Airflow) and S3-based data flows is a plus
- AWS fluency and CI/CD discipline are advantageous

Qualifications
- Relevant quantitative or technical degree preferred but not mandatory if experience demonstrates capability and expertise