Time Type:

Full time

Working Pattern:

Hybrid

​
Purpose of the Role

Working closely with the Data Science Manager, the holder will strengthen the data science capability of the business by delivering models and actionable insights that improve underwriting profitability and unlock automation and efficiency across teams. This is a true end to end role: the holder will own projects from problem framing through development and deployment, and will remain accountable for their models in life, monitoring performance and drift and judging when a model needs retraining or retirement. We are looking for someone who designs with a road to production mindset from the outset, with demonstrable experience of personally taking models into operational use.
Technical delivery is only half the job.

This role works directly and continuously with our commercial underwriters, from framing the problem through to adoption. The holder will translate underwriting requirements into data science solutions, build confidence in the results, and spend real time with underwriting teams understanding how each class operates.

Technically, we are looking for strong object oriented programming alongside ML, statistical and data analysis capability, and fluency in designing machine learning pipelines and deploying them (Azure in our case). MLOps and CI/CD experience is platform agnostic; the emphasis is on knowing the detail behind MLOPS patterns actively contribute to raise MLOps standards across the team.
Beyond delivery, the holder will act as an ambassador for the team, engaging across the business to capture and prioritise requirements and feeding these into a backlog that shapes the Data Science roadmap and sustains a consistent stream of value.

The right candidate will bring a mature mindset and genuine breadth across modelling, production grade programming, project management, commercial judgement along with the leadership and coaching skills to develop more junior colleagues. Close collaboration with Actuarial is expected from the outset, so experience of insurance pricing in a regulated environment, and comfort working alongside actuarial methodology, is a distinct advantage. Effective stakeholder management is core; the right candidate will spend time with teams to understand the business and how each class operates.

​​
Duties and Accountabilities

• Delivery of data science products

• Lead data science projects end to end, from problem framing and data preparation through development, deployment and ongoing monitoring in production.

• Work alongside the actuarial team to surface insights that drive performance (eg. Reserving).

• Apply data science techniques to automate manual processes across the business.

• Use generative AI to enrich insight and unlock new opportunities for the roadmap, deploying and maintaining these solutions through the same MLOps patterns applied to traditional models.

• Research, assess and integrate external data sources, working with data scientists and actuaries to establish their quality, value and fitness for use.

• Address the data quality issues that constrain modelling, including the matching of premium and claims for delegated business.

• Support the business with proactive analytics and insights, coordinating delivery with the Data Science and Data Analytics Manager.

• Engineering and MLOps standards

• Design, build and maintain machine learning pipelines in a cloud environment, applying sound software engineering practice.

• Set and raise the team's standards for version control, testing, CI/CD, model versioning and reproducibility.

• Own deployed models in life, monitoring performance and drift and acting on degradation before it reaches the business.

• Ensure models are documented and explainable to a standard appropriate for a regulated environment.

• Stakeholder engagement and requirements

• Work with technical and non technical stakeholders across the business to identify, document, analyse and prioritise requirements for data science products.

• Coordinate with IT and Data Engineering to shape the data foundations these products depend on, and contribute to best practice and operational procedures for data management.

• Produce clear deliverables and communicate findings and their limitations to audiences without a technical background.

• Team and capability building

• Coach data scientists and data analysts, through code review, pairing and technical mentoring, to build depth in modelling and production practice.

• Work with the Data Science Manager to upskill the team in new and emerging techniques, creating flexibility of resource while maintaining clear accountability.

• Contribute to the data science backlog and roadmap, advocating for projects with demonstrable value.

​​
Skills, Knowledge and Experience

Essential

• Well developed Python, written to production standard, with object-oriented design, testing and code review as normal practice.
• Demonstrable experience of personally taking models into production and supporting them in life.
• Machine learning across the standard toolkit (scikit learn, pandas, NumPy, statsmodels or equivalents), with sound judgement about model selection, validation and the limits of what the data supports.
• Software engineering fundamentals: version control, branching strategy, code review, automated testing, dependency and environment management.
• MLOps and CI/CD in practice, covering pipeline orchestration, model versioning, automated deployment, monitoring and retraining. Platform agnostic; what matters is knowing the patterns and being able to discuss them in technical detail.
• Cloud based machine learning delivery, ideally on Azure (Azure ML, Azure DevOps), with equivalent AWS or GCP experience considered.
• SQL and relational data modelling, with the ability to work efficiently against large datasets.
• Data wrangling and pre-processing, including the ambiguous, incomplete and inconsistent data typical of insurance.
• Statistical foundations sufficient to design sound experiments, quantify uncertainty and challenge conclusions that the data does not support.

Desirable

• Insurance experience, (specifically in Lloyds market) pricing or underwriting in a regulated environment, and comfort working alongside actuarial methodology.
• Practical experience deploying generative AI or LLM based solutions, including retrieval patterns, evaluation and cost, latency management and observability.
• Distributed processing frameworks such as PySpark.
• Data visualisation and reporting, for example Power BI.

Behaviours and ways of working

• Commercial judgement: chooses problems by the value they create, and knows when a simpler solution is the right one.
• Communicates clearly with non-technical audiences, including the limitations and uncertainty in their work, not only the results.
• Manages competing priorities and stakeholders without losing delivery focus.
• Works well across teams, and builds capability in others rather than concentrating knowledge in themselves.
• Sound ethical judgement and awareness of data privacy, fairness and regulatory obligations and the commercial impact of deliverables.

AEGIS Values

Fairness and respect

We make decisions considering the best interests of key stakeholders. We are direct and straightforward in our actions, working collaboratively to create a culture of fairness and respect.

Open and inclusive

We act with integrity, valuing diversity of thought and background. We take time to listen to the needs of our customers, stakeholders and colleagues working together to seek and share information.

Ambitious

We have a passion for success, aspiring to be recognised as best in class. We embrace new opportunities, encouraging innovation in pursuit of our goals.

Striving to be better

We strive to improve at all times, challenging complacency, being agile and adapting to change. We always seek to improve our customers’ experience with us.

Investing in people’s potential

We provide an environment where each employee can reach their personal potential. We encourage personal accountability for performance and individual ownership for growth and success.

AEGIS London is an equal opportunities employer and recognises the value of a diverse workforce in facilitating better decision making and business growth. We encourage a variety of differing views, perspectives and insights to create a collaborative working environment. Diversity and Inclusion are fundamental to our business and we encourage applications from all backgrounds recognising the diversity of society and our customers.

It’s important to us that you are able to perform at your best when applying for a role with AEGIS London. If there are any adjustments we can reasonably make to ensure that the process is accessible for you please telephone us on +44(0)20 7856 7856 or email [email protected]

As a business, we understand individual circumstances may differ and aim to be adaptable and to support flexible working practices. Talk to our recruitment team to understand how AEGIS London can help support you in reaching your full potential