AI Integration Architect

Role Summary

We are looking for a Senior AI Architect to design and deliver AI systems across Cotality's property intelligence platform. This is a hands‑on individual contributor role with broad scope spanning internal agentic systems that power property analytics and decision workflows, and external AI integration architecture that makes Cotality's data products consumable by AI agents, foundation model platforms, and enterprise developer ecosystems.

You will work across both layers, ensuring they are coherent, secure, and built to scale with the growth of our product portfolio. You will define the standards, build the foundational components, and be accountable for the architecture working in production under real client load.

• Design and build agentic AI systems, including multi‑agent frameworks, orchestration layers, memory and retrieval architectures, and tool‑based reasoning pipelines that operate against structured and unstructured property data.
• Own the external AI integration architecture, API gateway configuration, MCP server patterns, authentication and authorization flows, tool schema standards, and the reference architecture that product teams follow to expose their APIs as agent‑consumable tools.
• Pioneer Agent Experience (AX) design as a first‑class methodology for the organization analogous to UX or Developer Experience (DX). Treat AI agents as primary consumers of our systems and ensure that APIs, tool descriptions, and data outputs are optimized for LLM comprehension, context limits, and deterministic reasoning.
• Establish and enforce technical standards for how AI agents consume Cotality's data products, with a focus on tool description quality, input and output contracts, error handling patterns, and response metadata standards all evaluated through the lens of AX.
• Architect security and data provenance controls across the integration layer, including JWT claim schema design, defense‑in‑depth authorization patterns, audit logging, and response boundary enforcement.
• Design and implement observability and telemetry for AI systems to monitor token consumption, latency, error rates, prompt drift, LLM costs, and response quality in production.
• Establish CI/CD pipelines and evaluation frameworks for AI agents that measure accuracy, hallucination rates, and performance regressions before changes reach production.
• Optimize AI workload architecture by designing deployment strategies that decouple large model weights from application code, utilizing optimized base images and dynamic runtime mounting to maintain fast, reliable CI/CD pipelines.
• Scale inference and orchestration by architecting high‑throughput AI backends using specialized model servers such as vLLM or Triton on Kubernetes, with support for dynamic batching, streaming responses, and concurrent execution.
• Align application design with cloud economics by partnering with platform engineering to build cost‑aware AI systems, and designing agentic workflows that gracefully handle cold‑start latencies and infrastructure scaling events such as scale‑to‑zero or Spot instance evictions without dropping requests.
• Bring strong backend engineering practices to the AI layer, with a consistent track record of delivering production‑quality, maintainable code in cloud or containerized environments.

• Define the reference architecture and MCP server build patterns that product teams across the organization follow when exposing their APIs as agent‑consumable tools.
• Partner closely with data engineers, product managers, and domain experts in insurance and property risk to ensure that AI systems produce outputs that are accurate, traceable, and operationally meaningful.
• Review implementations, conduct architecture design reviews, and hold the technical quality bar as the connector portfolio grows and more teams contribute.
• Produce architecture decision records, technical standards, and reference documentation that engineering teams across the organization rely on to make consistent, well‑reasoned decisions.

• 7 to 10 years of software engineering or architecture experience,…