AVP, AI Engineering; Remote - EST

Position: AVP, AI Engineering (Remote - EST)
With a company culture rooted in collaboration, expertise and innovation, we aim to promote progress and inspire our clients, employees, investors and communities to achieve their greatest potential. Our work is the catalyst that helps others achieve their goals. In short, We Enable Possibility℠.

Job Summary

Strategic Analytics is a dynamic and growing team at Arch that drives innovation and transforms how the business operates. We build AI-first, agent-driven products that change how Arch underwrites, services, and learns from its book - combining frontier LLMs, multi-agent orchestration (MCP, A2A), retrieval-augmented generation, evaluation harnesses, and traditional ML where it still wins. Our mission is broad: agentic automation, decision support, AI-driven insights, and the platform engineering required to make all of it production-grade.

We have a strong track record of success (productionalizing dozens of gen AI products over the last 3 years) - we aim to continue scaling these efforts and are seeking an AVP AI Engineering Senior Agentic Software Engineer to lead the build-out of true multi-agent systems within Strategic Analytics. Reporting to the VP of AI & Automation, you will design and operate orchestrations where agents communicate directly with one another - not just sequential, hand-off-driven workflows - to deliver underwriting and claims automations end-to-end.

This role sits at the seam between automation engineering (AE) and implementation engineering (IE). You will own everything from the master agent down through worker agents, while partnering with IE on the orchestration entry points and any infrastructure-side connectors.

Key Responsibilities

* Design and build multi-agent orchestration patterns (master-orchestrator + specialized worker agents) using protocols such as MCP and A2A.

* Lead end-to-end delivery of agentic underwriting and claims automations, from prototype through production hardening.

* Partner with IE on orchestration entry-point design (e.g., Azure Function endpoints, master-agent gateways) so the AE/IE seam is clean and scalable.

* Lead offshore engineers and team members on agentic patterns, prompt engineering, and reliability practices.

* Establish coding, evaluation, and observability standards for agentic systems within the AI & Automation Center of Excellence.

* Translate business intent into working agentic systems - not just systems that compile, but systems that deliver measurable business outcomes.

Required Skills / Experience

* 7+ years of software engineering and/or automation engineering experience.

* 3+ years of people leadership experience

* Demonstrated experience building production grade multi-agent or agentic systems (not just POCs that did not ship).

* Strong track record of shipping AI, analytics, and/or automation products that have a financially measurable impact on the business

* Strong Python experience with API development, async patterns, and service orchestration.

* Hands-on experience with at least one agentic framework (e.g., Lang Graph, Claude Cowork / Agent SDK, OpenAI Agents, Auto Gen) and at least one inter-agent protocol (MCP, A2A).

* Resilient problem solving - comfortable with ambiguous problems and capable of breaking them into shippable increments.

* Strong written and verbal communication; able to operate across IE, AE, DS, and business stakeholders.

* Demonstrated, hands-on production experience (not POC-only) with the current AI stack: RAG, MCP / A2A or equivalent agent-to-agent protocols, agentic orchestration frameworks, and the ability to articulate where each is the right tool. Comfort scanning for and trialing new tooling as the space evolves.

Desired Skills / Experience

* Background that includes a stretch of "real" software engineering before moving into analytics or AI (e.g., SWE for a few years, then a master's in analytics, then 5+ years in the AI/ML space).

* P&C insurance domain familiarity - underwriting, claims, or submission lifecycle.

* Experience with retrieval-augmented generation (RAG), evaluation harnesses, and structured-output patterns.

* Cloud experience in Azure (preferred for our stack) and/or AWS;…