Overview:
In this role you own LLM-powered systems and agentic workflows within CCC’s insurance-focused platform, guiding the end-to-end product area from problem framing to production. You define evals, guardrails, and privacy baselines to ensure safe, scalable deployments as models evolve. You will work with product managers and data scientists to shape durable, multi-tenant architectures and maintain auditability. The role emphasizes robust design, security, and cross-functional collaboration to deliver trusted AI-powered experiences. Compensation / Benefits 401K Match Paid time off Annual Incentive Plan Performance Bonus Comprehensive health insurance Adoption Assistance Tuition Reimbursement
Responsibilities:
Own the delivery of LLM-powered systems in Python and TypeScript from ambiguous customer problems to production operation Define evals and guardrails strategy for your area, including release gates and production monitoring Refine agentic coding tool usage by sharpening specs, hooks, and permission policies, with code review standards Design NLP pipelines to convert unstructured claim data into actionable structured data, selecting semantic search, classical NLP, or LLM where appropriate Build and operate durable, event-driven, multi-tenant systems with deterministic workflows and well-defined failure/replay semantics Own tracing and audit for your area to reconstruct and cite model/agent outputs from sources Establish privacy and security baselines like tenant isolation and least-privilege access for agents and tools Author and review technical design documents; mentor engineers and participate in on-call rotations Collaborate with teams to align on scope and drive cross-team delivery
Key requirements:
6+ years of professional software engineering experience with an architected, shipped LLM-powered system Proficiency in Python and TypeScript with strong conventions, typing, async model, and test strategy Experience defining evals/guardrails strategy and production correction workflows Experience constraining and validating agentic coding tool outputs; daily use of tools like Claude Code, Cursor, GitHub Copilot Designs for pipelines blending LLMs, semantic search, and classical NLP; defensible retrieval design Durable, event-driven systems with deterministic orchestration, compensation actions, idempotency, backpressure, and event-sourced replay Data modeling across PostgreSQL, MongoDB, Neo4j; knowledge of partition keys, consistency, tenant isolation Systems architecture on Azure or AWS with CI/CD and infrastructure as code; experience with Temporal or Durable Task Scheduler Technical design documents authored across multiple teams; strong written and cross-functional communication skills Customer problem framing and trade-off communication to technical and non-technical audiences Clear, customer-focused communication Mentorship and collaboration across teams Bias toward action and owning outcomes Python TypeScript LLM-powered systems