Senior Data Scientist at Discovery Limited

Role Purpose

• You are a core member of Vitality Actuarial and Data Science (VADS), responsible for turning data into intelligent products that change how members experience Vitality and producing insights that change the business. You work at the intersection of behavioural science, health data, financial services and applied AI and you own the journey from a raw idea to something live in members' hands.
• Your primary responsibility is to find and extract new value from data that others haven't seen yet. You do this by being intensely curious about the business, the members, the data estate, and the frontier of what AI can now do. You go looking for problems. You prototype fast. You ship. Then you find the next one.
• You are the team's connective tissue to the outside AI world; the person who knows what shipped last week, what's hype, and what's quietly changing everything and you bring that back with a point of view, not a link dump.

Key Responsibilities

Building & Shipping Intelligent Products

• Own ML products end-to-end  from data exploration and problem framing through to production deployment on GCP (Vertex AI, BigQuery) and/or Azure Databricks. You don't hand things off at the notebook stage.
• Deploy and operationalise LLMs and ML models that drive personalisation, engagement and intelligent recommendations for millions of Vitality members.
• Write production-grade code, not just experiments. You care about MLOps, monitoring, and what happens to your model on day 90, not just day one.
• Partner with engineering, platform and data teams to strengthen Vitality's AI platform and data products. You hold the technical bar on what "good" looks like.

Extracting New Value from Data

• Go looking for value proactively surface use cases that unlock member impact, business value or entirely new products. Nobody needs to hand you a brief.
• Interrogate the data estate. Know what we have, what we don't, what's underused, and what external data would change the game. Come with proposals.
• Prototype fast, kill fast, scale what works. You're comfortable running multiple bets and being honest about what's not landing.
• Frame the "so what." Every model, every insight you can articulate the decision it changes and the value it creates.

Being the Team's Signal on AI

• Stay obsessively current on what's happening in AI, ML and LLMs. Not as a hobby, but as a professional discipline.
• Separate signal from noise and bring back a curated, opinionated view of what actually matters for Vitality.
• Run experiments with new tools and techniques before anyone asks. Come with evidence, not opinions.

Raising the Team Around You

• Upskill junior members on modelling, engineering discipline and how to think about problems, not just solve them.
• Demystify AI for actuarial, business and executive stakeholders. Translate without dumbing down.
• Translate complex technical work into clear, compelling narratives for Vitality Exco, the Discovery Board and Group Exco.

Core Skill Development

• Applied machine learning and LLM engineering — model development, fine-tuning, RAG, evaluation, prompt engineering, and productionisation.
• MLOps and production discipline — CI/CD for models, monitoring, drift detection, versioning, cost management.
• Cloud-native data science — GCP (Vertex AI, BigQuery) and/or Azure Databricks fluency.
• Data fluency across the Vitality estate — knowing where data lives, how it flows, and where the untapped value sits.
• Business translation — converting behavioural, actuarial and commercial questions into precise technical problems.
• Executive storytelling — framing technical work so senior leaders can make decisions on it.
• Behavioural science literacy — enough to design products that actually change what members do.

Who You Are

• 5–8+ years in data science, ML, or a closely related quantitative field.
• A degree in a quantitative discipline (actuarial science, statistics, computer science, data science, maths, physics, or similar). Postgraduate work is a plus.
• Proven track record of taking models from prototype to production; you understand MLOps, not just model accuracy.
• Strong Python and SQL are expected. Experience with GCP (Vertex AI, BigQuery), Databricks, LLM-based applications and hands-on data engineering is beneficial.
• Genuine curiosity about behavioural science, health, wellness, or insurance — not just the tech.