Data Scientist - Agentic AI (hybrid)

Allstate Allstate · Insurance · Belfast, United Kingdom +1

Data Scientist role focused on developing and maintaining data-driven agentic solutions using LLMs and cloud-based AI. The role involves the full lifecycle from data collection to model/agent development, deployment, and monitoring, with a focus on building foundational experience in AI agent design and scalable AI systems.

What you'd actually do

  1. Learns and applies LLMs, programming languages, and tools used by the team to improve efficiency and solve business problems. Assists in exploring new data sources, research, and models under direction from senior team members.
  2. Supports project planning efforts by breaking down AI/predictive modeling and development tasks into manageable subtasks, tracking progress and meeting assigned deadlines. Executes analytical and modeling tasks as part of a larger project team.
  3. Applies standard best practices to develop statistical and machine learning techniques to build models that address business needs and improve data quality and decision making.
  4. Reviews and discusses AI/modeling techniques and results with peers and mentors, clearly communicating findings and incorporating feedback.
  5. Contributes to stakeholder communications by preparing summaries, documentation, and analysis to support senior team members in influencing business partners and leaders.

Skills

Required

  • Python
  • data analysis
  • modelling
  • building production ready solutions
  • R
  • SQL
  • TensorFlow
  • PyTorch
  • CI/CD pipelines
  • containerization (Docker)
  • observability concepts
  • cloud security fundamentals
  • AI agent development frameworks (e.g., Azure AF, AWS Strands, Google ADK, LangGraph, OpenAI Agents SDK)

Nice to have

  • risk management
  • AI governance concepts
  • AWS Bedrock
  • Azure Foundry
  • Infrastructure as Code (IaC)
  • no-code/low-code agents using M365
  • degree in a quantitative field (mathematics, physics, computer science) with classification of 2.1 or higher

What the JD emphasized

  • legal right to work in the UK
  • minimum 3 years of hands on, practical experience using Python for data analysis, modelling, and building production ready solutions
  • Working knowledge or coursework exposure to CI/CD pipelines, containerization (Docker), observability concepts, and cloud security fundamentals
  • Introductory experience or strong interest in AI agent development using at least one framework

Other signals

  • AI agent development
  • LLMs
  • production environments
  • cloud-based AI solutions