Lead Applied Scientist

AXIS (AXIS Capital)•Halifax, NS

3d•$175,000 - $200,000•Hybrid

About The Position

This is your opportunity to join AXIS Capital – a trusted global provider of specialty lines insurance and reinsurance. We stand apart for our outstanding client service, intelligent risk taking and superior risk adjusted returns for our shareholders. We also proudly maintain an entrepreneurial, disciplined and ethical corporate culture. As a member of AXIS, you join a team that is among the best in the industry. At AXIS, we believe that we are only as strong as our people. We strive to create an inclusive and welcoming culture where employees of all backgrounds and from all walks of life feel comfortable and empowered to be themselves. This means that we bring our whole selves to work. All qualified applicants will receive consideration for employment without regard to any protected characteristic, including age, color, disability, ethnicity, gender identity, marital status, national origin, pregnancy, race, religion, sex, sexual orientation, veteran status, or any basis prohibited by the laws that govern its operations. How does this role contribute to our collective success? Data and analytics are of critical importance for AXIS. We turn data into information so the business can make decisions with confidence, identify opportunities early, and operate more efficiently. The Lead Applied Scientist is the senior scientific authority in the Data Science and AI Delivery team and is responsible for the quality of the AI the team produces, much of which now uses large language models and agentic approaches. The role leads this work from the point a business problem is framed, through the choice of approach and the build itself, to how the resulting solution is evaluated, validated and monitored in production. Working alongside the engineering lead, it sets the scientific standards that give the business confidence in what the team delivers.

Requirements

Strong, hands-on expertise with large language models and generative AI, including prompt engineering, retrieval augmented generation, fine tuning, and the design and evaluation of agentic systems that use tools and operate over multiple steps.
Strong programming ability in Python and fluency with the modern AI and data science tooling, including frameworks for building with large language models alongside libraries such as pandas, NumPy and scikit-learn.
Strong mathematical and statistical foundations, including probability, statistics, linear algebra and optimization, and the ability to reason rigorously about uncertainty, error and model behavior.
A sound grounding in machine learning, including model validation and experimental design, applied where it is the right tool for the problem.
Demonstrable experience designing evaluation and validation frameworks for AI systems, including methods for measuring the quality of large language model, retrieval augmented generation and agent outputs.
A strong track record of building AI solutions that solve real business problems and of taking them from prototype to production.
Practical experience with Databricks, MLflow and Spark based data processing.
Working knowledge of responsible AI methods, including bias and fairness testing, explainability techniques, and model risk.
The ability to explain complex technical concepts clearly to non-technical and senior audiences.
A track record of setting technical standards and developing other scientists and engineers.

Nice To Haves

Experience with agent frameworks and orchestration tools for building multi-step, tool-using AI systems.
Experience in the insurance or reinsurance industry, with an understanding of underwriting, claims or actuarial data.
A postgraduate qualification in a quantitative or computational subject such as computer science, statistics, mathematics or a related field.

Responsibilities

Lead the team’s approach to solving business problems through advanced analytics and AI, guide how solutions are designed and built, and continue to contribute directly to technical work such as model development, experimentation, code review and solution architecture.
Lead the design and development of the team’s AI and ML solutions, choosing the right approach for each problem, with large language models and agentic AI increasingly central to the work.
Bring mathematical and statistical rigor to how problems are framed, how uncertainty is handled, and how the team judges whether a solution is good enough.
Build and evaluate solutions to the most demanding problems the team takes on, review the work of others to keep standards high across the delivery scrums, and act as the point of escalation for difficult technical decisions.
Hold sign-off on scientific approach and solution quality.
Leading the design and development of the team’s AI/ML solutions, and framing business problems so that the right approach can be chosen and the result measured against clear success criteria.
Leading the design of generative and agentic AI solutions, including prompting, retrieval augmented generation, tool use and multi-step agent workflows, and the techniques needed to make them accurate and reliable.
Selecting the right approach for each problem, with large language models and agentic AI to the fore, and drawing on deep learning, machine learning and statistical methods where they are the better fit.
Owning the evaluation and accuracy methodology for the team’s models, agents and AI systems, including the metrics, test sets and acceptance thresholds that govern performance, with proper treatment of uncertainty and statistical significance, and the monitoring needed to detect drift in production.
Leading the team’s responsible AI work, including bias and fairness testing, explainability, and validation of model and agent behavior against regulatory expectations.
Solving the team’s most challenging problems, such as extracting information from unstructured documents, automating expert workflows with agents, optimization, forecasting, and portfolio and claims analytics.
Owning the scientific design decisions on each solution, including choice of approach, model and method selection, and evaluation strategy, while working closely with the engineering lead on architecture and deployment.
Writing production quality code and building solutions directly, particularly on novel or higher risk work.
Setting the experimentation, evaluation, coding and documentation standards for the team, and raising them through code review, pairing and technical mentoring.
Explaining methods, results and trade-offs clearly to business stakeholders, model risk, and governance forums.