Senior AI Engineer, Agentic Apps & Personal Agents
About the Client
The client is a commercial insurance technology company and a subsidiary of a major national insurance group. It runs a digital exchange connecting brokers, agents and carriers, and has already put a production agentic AI platform in place that automates the commercial insurance journey from quote request to bind. This role builds the next layer on top of that platform.
The Role
You’ll embed directly in the client’s AI engineering team, joining sprints and standups and working Jira tickets in a flat, collaborative structure. You’ll work closely with the AVP of AI & Data Science. You’ll build AI applications for the major frontier-model ecosystems and a new set of personal AI agents, take them into production, and then support what you’ve built.
What You’ll Do
Build and ship AI apps for ChatGPT, Claude, Grok and Gemini, including each platform’s SDKs, hosting, packaging and marketplace or publishing requirements.
Build personal AI agents similar to Dot, Muse and Duet, with memory, context and proactive action.
Build MCP servers and integrations that expose the existing agentic platform’s capabilities to frontier-model apps and agents.
Build the APIs and full-stack components these applications need.
Implement guardrails, evaluation and observability, and partner with QA on golden-set accuracy testing.
Own work end to end, from design through production, including the edge cases of each model provider.
Over time, contribute to and extend the core agentic workflow, and maintain and support everything you build.
Must-Have
Production AI experience: 2–3+ years of hands-on, production-grade LLM and agentic application work. Prototypes and side projects alone won’t qualify.
Engineering background: 5+ years of software engineering overall, with full-stack experience (Python and/or TypeScript/Node, a modern front-end framework, and REST/GraphQL APIs).
Agent frameworks: Deep experience with frameworks such as LangGraph, LangChain or CrewAI.
Provider SDKs: Expert with the OpenAI SDK (including the Agents SDK), plus working knowledge of at least one other frontier provider’s API (Anthropic, xAI or Google).
MCP: Hands-on experience building MCP servers or integrations.
Agent design: Strong command of tool calling, structured outputs, RAG, memory, multi-agent orchestration, guardrails and evaluation.
Collaboration: Excellent communication skills and comfort working embedded in a core team rather than in a silo.
Strongly Preferred
Has published an app or connector to the ChatGPT apps ecosystem, the Claude connector directory or a similar marketplace
Has built a consumer-facing or personal AI agent
Familiar with emerging agent interoperability protocols such as A2A
Has experience with eval and observability tooling such as LangSmith, plus MLOps practices
Has worked in insurance, fintech or another regulated industry
Not a Fit
Traditional developers whose AI experience is limited to recent tutorials or experiments
“Vibe coding” without engineering rigor, testing or production ownership
Benefits we offer
Access to e-Learning platforms.
Amazing people-oriented organizational culture
Working from anywhere
Challenging projects using the latest technologies with clients from the US.