Data Scientist - Decision Intelligence & Agentic Solutions

Role Purpose

• Turn analytical intelligence into decisions, interventions and agentic workflows by combining strong Data Science foundations with decision design, workflow reasoning and practical GenAI / agentic capabilities.
• Apply Data Science methods to prediction, prioritization, segmentation, experimentation and optimization, then translate the outputs into decision logic that business workflows can realistically use.
• Define what downstream agents or applications need beyond a model score: context, explanations, constraints, eligibility, confidence, action options and escalation rules.
• Determine where decisions should be deterministic, model-driven, LLM-assisted or human-approved; design safe autonomy boundaries and guardrails.
• Prototype and evaluate agentic workflows using tools, APIs, retrieval, structured outputs and function calling where they add genuine value.
• Measure workflow / agent performance alongside model performance and create feedback loops linking interventions to customer and business outcomes.

Job Description

Strategic Planning

• Translate priority business workflows into decision-intelligence and agentic use cases with clear value hypotheses.
• Identify where deterministic rules, predictive models, LLMs, agents or human approval should be used - and where they should not.
• Define reusable decision patterns and analytical interfaces that can scale across multiple business domains.

Operational

• Build analytical components for prediction, prioritization, treatment selection and optimization, then expose them as usable decision inputs.
• Prototype agentic workflows using retrieval, tools, APIs, structured outputs and function calling where appropriate.
• Design guardrails, confidence thresholds, escalation paths, auditability, evaluation and monitoring for production workflows.

Cross-Functional Delivery

• Partner with AI CoE to consume and operationalize domain models; partner with Agent Engineering and Workflow Integration teams for production implementation.
• Work with Business Translators / Value Assurance to ensure decision products solve the business problem and create measurable impact.
• Create outcome and feedback loops that improve model, workflow and agent behavior over time.

External Relation

• Evaluate relevant GenAI, agentic, workflow and integration technologies based on measurable delivery value rather than novelty.
• Work with technology partners when specialized capability is needed while protecting internal ownership of critical IP and operating knowledge.
• Stay current on agent evaluation, orchestration, safety and production practices.

Qualification:

• Minimum bachelor’s degree in STEM, Statistics, Mathematics, Computer Science, Engineering, Economics or another quantitatively rigorous discipline.

• Experience:

• Minimum 4-8 years across Data Science, applied ML, decision intelligence or AI product development, with evidence of translating analytics into operational actions. Hands-on GenAI / agentic experience is highly desirable.

Skills:

• Strong Python and SQL capability with solid machine learning, statistics, feature engineering and experimentation fundamentals.
• Experience designing prioritization, next-best-action, treatment selection, recommendation, optimization or decision-support solutions.
• Ability to separate prediction, business policy and workflow orchestration, and to reason about expected value and intervention economics.
• Practical GenAI capability including foundation models, prompting, RAG, structured outputs, function/tool calling and agent orchestration concepts.
• Experience designing API / tool interfaces or working closely with engineering teams to operationalize analytical outputs.
• Understanding of guardrails, human-in-the-loop controls, fallback paths, permissions, evaluation and observability for agentic systems.
• Ability to design experiments and feedback loops that measure intervention and workflow outcomes rather than model metrics alone.
• Strong business translation and communication across technical, product and operational stakeholders.