Director, AI Native Analytics

Locations: Radnor, PA; Charlotte, NC; Fort Wayne, IN; Greensboro, NC. Work arrangement: hybrid (3 days per week in a Lincoln office). Relocation assistance is not available. Requisition #76794.

About the role

Lincoln Financial Group is building an AI Native Intelligence platform that turns enterprise data into a conversational, agentic layer the business can query directly — and that continuously finds patterns, trends, and anomalies in that data on its own. The platform's architecture spans an AWS-native ingestion and extraction pipeline, a governed semantic layer and business knowledge graph, and an intelligence layer that answers questions in plain language and surfaces what's changing, instead of relying on static reports someone has to remember to check.

We're looking for a Lead, AI Native Analytics to be the hands-on builder of that conversational and analytical experience: the person who actually designs and ships the Bedrock agents, retrieval pipelines, pattern-detection logic, and natural-language-to-analytics capabilities that make the platform usable. This is an individual-contributor-heavy, deeply technical role — you'll spend most of your time in the architecture, the prompts, and the code, while partnering closely with the Senior Director on direction and with business stakeholders on what “useful” actually looks like.

Responsibilities

• Design and build the conversational AI experience end to end: Bedrock Agents (via AgentCore or MCP/API), retrieval-augmented generation over Bedrock Knowledge Bases / OpenSearch Serverless, natural-language-to-SQL over Athena, and relationship queries over a knowledge graph (Amazon Neptune / GraphRAG).
• Build capabilities that continuously detect patterns, trends, and anomalies across enterprise data — surfacing what's notable proactively, rather than waiting for someone to know what question to ask.
• Build and tune the AI extraction pipeline that turns raw unstructured data (documents, logs, notes, communications) into structured, query-ready facts — using Amazon Comprehend and/or LLM-based extraction via the Bedrock Converse API (Lambda, Step Functions, DynamoDB for state).
• Own prompt engineering and agent tool design: system instructions, tool/action-group definitions, structured-output schemas, and model selection/tiering to balance accuracy, latency, and cost.
• Integrate the conversational layer with the platform's semantic layer and business ontology (KPI and metrics definitions, business rules glossary) so agent answers stay consistent with how the business already measures itself — never computing an ad hoc number that disagrees with a dashboard.
• Build evaluation and quality processes for LLM-based systems: golden question sets, retrieval/accuracy scoring, hallucination and abstention testing, and regression checks before any change ships to users.
• Implement guardrails in partnership with security/compliance: PII detection and redaction, Bedrock Guardrails configuration, access-controlled data zones, and audit logging for every AI-generated answer.
• Partner directly with business stakeholders across the enterprise to translate real questions into working conversational and pattern-detection features, and iterate quickly based on usage and accuracy feedback.
• Mentor and provide technical direction to engineers contributing to the conversational AI workstream, and represent this workstream in architecture reviews led by the Senior Director.
• Stay current with the fast-moving AWS AI landscape — Bedrock model and AgentCore updates, new retrieval and evaluation tooling — and bring concrete recommendations back to the team.

Requirements

• 8+ years of experience in Data Analytics / Business Intelligence with at least 2+ years building production LLM-based Analytics systems (not just prototypes or proofs of concept).
• Hands-on, production-level expertise with the AWS AI/data stack: Amazon Bedrock (Knowledge Bases, Agents/AgentCore), AWS Lambda, Step Functions, Amazon S3, AWS Glue, and Amazon Athena; working experience with a vector store (OpenSearch Serverless or similar).
• Strong, practical prompt-engineering and LLM application development skills: structured extraction, tool-calling/agentic patterns, retrieval design, and iterative evaluation against real questions.
• Solid understanding of semantic layers, business ontologies, or metrics layers, and how to ground LLM outputs in governed, consistent business definitions rather than ad hoc computation.
• Experience building or applying pattern-, trend-, or anomaly-detection techniques over enterprise data (statistical, ML-based, or LLM-assisted), and turning what's found into something a business user can act on.
• Strong proficiency in Python and/or TypeScript/Node.js for building extraction pipelines, agent tools, and integration code.
• Working knowledge of SQL and cloud analytics engines (Athena, Redshift, or similar), sufficient to design and validate natural-language-to-SQL experiences.
• Clear communicator who can demo a working system to engineers and explain its behavior and limitations to non-technical business stakeholders with equal clarity.
• Bachelor's degree in Computer Science, Engineering, or a related field, or equivalent practical experience.

Preferred qualifications

• Experience with GraphRAG or other knowledge-graph-grounded retrieval patterns, and with a graph database such as Amazon Neptune.
• Experience with a semantic-layer/metrics-layer tool (e.g., Cube.dev) used to serve consistent metrics to both BI and AI consumers.
• Experience with NLP/document-processing pipelines (e.g., Amazon Comprehend) for extracting structure from varied unstructured enterprise data.
• Experience in insurance, financial services, or another regulated industry, with a working understanding of PII handling and data privacy/compliance considerations.
• Experience configuring LLM safety/evaluation tooling such as Bedrock Guardrails, or building your own evaluation harness for generative AI systems.
• Familiarity with a range of enterprise data domains (financial, operational, customer, or workforce data) within a large organization.

Benefits

• Clearly defined career tracks and job levels, along with associated behaviors for each of Lincoln's core values and leadership attributes
• Leadership development and virtual training opportunities
• PTO/parental leave
• Competitive 401K and employee benefits
• Free financial counseling, health coaching and employee assistance program
• Tuition assistance program
• Work arrangements that work for you
• Effective productivity/technology tools and training

Pay

The pay range for this position is with anticipated pay for new hires between the minimum and midpoint of the range and could vary above and below the listed range as permitted by applicable law. Pay is based on non-discriminatory factors including but not limited to work experience, education, location, licensure requirements, proficiency and qualifications required for the role. The base pay is just one component of Lincoln's total rewards package for employees. In addition, the role may be eligible for the Annual Incentive Program, which is discretionary and based on the performance of the company, business unit and individual. Other rewards may include long-term incentives, sales incentives and Lincoln's standard benefits package.

Schedule

Hybrid: employee will work 3 days a week in a Lincoln office.

Application deadline

Applications for this position will be accepted through January 30, 2026, subject to earlier closure due to applicant volume.

About the company

Lincoln Financial (NYSE: LNC) helps people to confidently plan for their version of a successful future. We focus on identifying a