AI Agent Developer

Hagerty

1d•Hybrid

About The Position

As an AI Agent Developer, you will design, build, and continuously improve AI agents that handle real member interactions across the full customer lifecycle: from quoting and binding coverage, to policy service, to first notice of loss and claims. You will own the end-to-end agent development lifecycle on the Sierra platform, scoping agent behavior, designing conversation flows, integrating with Hagerty's systems, evaluating quality, and iterating relentlessly in production until each agent is the best version of itself. This is a senior individual contributor role. You are expected to operate with autonomy, partner directly with business and technology stakeholders, bring engineering discipline to everything you ship, and drive outcomes that are measurable in member satisfaction, resolution rates, and reduced manual effort.

Requirements

Hands-on experience building or deploying AI/LLM agents in production — including prompt engineering, eval frameworks, RAG, and agent tooling.
Strong technical aptitude: comfortable reading API documentation, understanding data models and system architecture, and engaging in engineering tradeoff discussions.
Experience with low-code/no-code or declarative platforms and the ability to ramp quickly on new tooling.
Proficiency consuming and integrating APIs, including authentication patterns and working with service accounts.
A degree in Computer Science, Engineering, Mathematics, a related technical field, or equivalent practical experience or comparable experience at the intersection of technology and business.
Experience applying engineering-quality practices to no-code/low-code development: automated testing where possible, regression coverage for workflows, monitoring and alerting for critical paths.
Experience with CI/CD, release management, and deployment promotion practices (dev/test/prod), including versioning and rollback awareness.
Strong security posture: least-privilege access, secrets management, auditability, and compliant handling of sensitive member data.
Ability to design solutions with clear inputs, outputs, and predictable behavior — built to be maintained and observed over time.
Proven ability to own complex, high-visibility projects end-to-end — from scoping and stakeholder alignment through production launch and continuous improvement.
Excellent verbal and written communication skills; able to explain technical concepts clearly to non-technical partners and drive alignment across teams.
Strong analytical and problem-solving skills — able to identify risks, develop mitigation strategies, and handle unexpected challenges autonomously.
Experience working at the intersection of technology and business operations, with a track record of delivering measurable outcomes.

Nice To Haves

Experience in insurance, financial services, or another regulated industry.
Familiarity with telephony platforms and voice agent development.
Background in conversational design, customer experience, or contact center operations.
Familiarity with public company requirements, including Sarbanes Oxley and key regulations, if applicable. For SOX compliant roles, responsible for designing, executing, and documenting internal controls where they have been identified as owners to prevent errors in financial reporting, processes, and business operations. Including attestation to the completeness, accuracy, and compliance of all financial reporting data, where applicable.

Responsibilities

Own the Agent Development Lifecycle: Take agents from initial scope through conversation design, tooling, evaluation, launch, and continuous iteration in production.
Translate business goals and member needs into agent behaviors that handle complex, high-volume interactions across quoting, policy service, and claims.
Define evaluation criteria for each agent, then use those metrics to drive improvement in resolution rates and member satisfaction.
Coordinate across technical and non-technical stakeholders throughout the full build-and-launch process.
Design and Build High-Quality Agents: Design and build conversational AI agents on the Sierra platform — expressing goals, procedural knowledge, and deterministic guardrails for regulated or high-stakes interactions.
Develop and refine conversational flows, rewrite prompts, and run simulations to tune agent behavior against real member scenarios — edge cases, tone, brand voice, and accuracy all matter.
Integrate agents with Hagerty's internal systems and APIs (policy administration, quoting engines, claims platforms) using appropriate authentication patterns and service accounts.
Build and maintain immutable agent release snapshots — including model version dependencies, knowledge bases, and prompts — following established release and engineering standards.
Apply engineering discipline in a no-code/low-code and declarative platforms: practical logging, error handling, monitoring, automated testing, and regression coverage for critical agent workflows.
Validate data accuracy in agent outputs, build reconciliation and anomaly checks where outcomes touch financial or policy data, and raise issues early.
Drive Continuous Quality and Improvement: Establish and run structured quality assurance processes — regularly auditing live conversations, annotating agent behavior, and converting findings into regression tests.
Use annotated conversations to build a growing test suite that prevents regressions across model upgrades, prompt changes, and platform updates.
Monitor agent performance in production, respond to incidents, and iterate rapidly based on usage signals, member feedback, and business outcomes.
Participate in on-call production support, sharing responsibility for incident response and agent reliability.
Shape How Hagerty Builds with AI: Act as a trusted internal advisor on AI agent strategy — contributing data-driven insights on what's working, where agents can expand, and how to push the platform further.
Identify and champion new agent use cases across the member lifecycle as the technology and Hagerty's capabilities evolve.
Contribute to team engineering standards, reusable integration patterns, and secure-by-default guardrails for agent development.