Senior AI Engineer (CCC Auto / Agentic)

EvolutionIQ•Chicago, IL

1d•$180,000 - $225,000

About The Position

As a Senior AI Engineer, you will own LLM-powered systems and agentic workflows for CCC customers. These systems automate routine parts of insurance claims and guide customers through the rest, requiring accuracy that can be measured and defended as models change and usage grows. Agents operate as steps within deterministic workflows, with privacy and security treated as design baselines. You will own a product area end to end, define the evaluation criteria it must meet before release, and use agentic coding tools while applying engineering judgment in architecture, code quality, security, and debugging. You will take systems from ambiguous customer problems through scoping with product managers and data scientists, delivery, and production operation. The role includes designing NLP pipelines and durable event-driven systems, establishing evaluation, observability, privacy, and security practices, authoring and reviewing technical designs and code, mentoring engineers, and participating in a shared on-call rotation.

Requirements

At least 6 years of professional software engineering experience.
Experience architecting and shipping an LLM-powered system to external customers and carrying it through at least one model or prompt migration.
Strong Python and TypeScript skills, including setting conventions for asynchronous programming, strict typing, and testing in a shared codebase, and profiling and fixing a performance problem.
Experience setting evaluation and guardrail strategies for a product area, including human-approval requirements and feeding production corrections back into evaluations.
Daily production use of an agentic coding tool such as Claude Code, Cursor, or GitHub Copilot; ability to explain how you constrain its context, verify output, and reject unsuitable changes.
Experience designing pipelines that combine LLMs, semantic search, and classical NLP, including defensible retrieval design and an example where a classical technique outperformed an LLM on cost, latency, or reliability.
Experience designing durable, event-driven systems with deterministic orchestration around non-deterministic model steps, including compensating actions, idempotency, backpressure, and event-sourced replay.
Data modeling and development across relational, document, and graph databases such as PostgreSQL, MongoDB, and Neo4j, including decisions about database fit, partition keys, consistency, and tenant isolation.
Experience architecting and operating systems on Azure or AWS, with CI/CD, infrastructure as code, and hands-on experience with a durable workflow engine such as Temporal or Durable Task Scheduler.
Experience authoring technical design documents for systems spanning multiple teams, obtaining feedback and buy-in before implementation, and having those teams build to the designs.
Clear written and verbal communication with technical and non-technical audiences, including eliciting customer needs and explaining trade-offs.

Nice To Haves

React and TypeScript experience delivering full-stack features, especially chat, streaming, and citation-heavy AI experiences.
Experience with agent frameworks such as LangGraph, PydanticAI, or Microsoft Agent Framework, and agent tool-access methods such as MCP, A2A, command-line tools, and microservices.
Experience with LLM observability and evaluation tools such as LangSmith, Arize, or Braintrust, and OpenTelemetry-based tracing.
Experience designing A/B tests for AI features.
Exposure to insurance, claims, or automotive repair domains.

Responsibilities

Own delivery and production operation of LLM-powered systems in Python and TypeScript, from customer problem scoping with product managers and data scientists through deployment.
Set evaluation and guardrail strategies, including release gates, production signals, and simulation of changes against historical claims.
Improve the team’s use of agentic coding tools through specifications, instruction files, hooks, and permission policies; review generated code to the team’s standards.
Design NLP pipelines that convert unstructured claim data into structured data, choosing appropriate uses for semantic search, classical NLP, and LLMs.
Build and operate durable, event-driven, multi-tenant systems with agents as steps in deterministic workflows; design failure handling, replay semantics, and cost profiles.
Own tracing and auditing so model calls and agent steps can be reconstructed and recommendations cited to their sources.
Establish privacy and security baselines, including tenant isolation, least-privilege access, redaction, retention, and rules for data shared with model providers.
Author design documents for complex systems, obtain feedback and buy-in, review designs and code, and mentor engineers.
Participate in a shared on-call rotation for owned systems.