Senior Software Engineer, AI Platform

At Kinetic, a single, powerful mission unites us:
Creating the best Workers’ Comp experience. Our culture is the engine that drives this mission forward—a culture built on togetherness, excellence, accountability, and a deep commitment to our customers.

We’re transforming workers’ comp with technology-driven solutions for injury prediction, prevention, and management—helping employers create safer, healthier workplaces.

Our platform is built for labor intensive industries including parcel delivery, manufacturing, and wholesale/warehousing operations. Our tech-driven program speeds up return-to-work and offers generative AI for claims management. Through these and other products and services, we're redefining what workers' comp can do.

We’re a remote-first team headquartered in New York City, growing fast—10x in the past two years—and driven by accountability, continuous improvement, and collaboration. If you're excited to build a career with impact, we’d love to meet you.

Kinetic has been named to:
Business Insurance's Best Place to Work in Insurance 2025 & 2026.

As a Senior Software Engineer on our AI Platform team, you'll build and enhance the platform every Kinetic AI agent runs on. We are building agents across the insurance lifecycle, from broker submissions to underwriting to claims. You'll own the orchestration layer, tools, evaluation infrastructure, and guardrails behind every agent. With that platform in place, we can release an agent safely, measure whether it improved a real business result, and quickly undo a change that doesn't work.

This is a hands‑on, senior individual contributor role. You'll write production code every day, set technical direction for agent infrastructure, and partner with our product and platform development teams as they build on this AI platform. You'll report to our Tech Lead - Data & AI.

• Build and operate Kinetic's agent orchestration layer: model routing, tools, memory, permissions, and guardrails, kept portable across frameworks and model providers.
• Manage agents like code, with prompts, tools, and model versions shipped through CI/CD, tracked for cost and latency, and easy to roll back.
• Build MCP servers and other tools that give agents secure access to Salesforce, Athena, and claims data, with PHI and PII redaction and full audit logs.
• Build the evaluation harness and rollout gates that test every change against golden datasets before it reaches production.
• Build human-in-the-loop checkpoints for critical decisions that always need a person.
• Replatform or retire the one‑off AI scripts and workflows already in use.
• Set standards, review agent designs and code, and mentor engineers building agents across Kinetic.
• Evaluate when to use or not use AI, and when to buy vs. build new systems and components.

• Bachelor's degree in Computer Science or a related field, and 6+ years of software engineering experience.
• 2+ years building LLM-powered systems in production, not just demos.
• Strong Python on serverless (AWS preferred), plus infrastructure as code and CI/CD.
• Hands‑on experience building multi‑agent systems with CrewAI or other agent frameworks such as Lang Graph, OpenAI Agents SDK, Google ADK, or AWS Bedrock Agent Core, including tool calling, structured outputs, RAG, and MCP.
• Built evaluation and observability for LLM systems, such as golden datasets, automated graders, and tracing.
• Clear communicator who can explain probabilistic systems to business partners.

• Experience at an insurtech, MGA, or carrier, or in another regulated industry such as healthcare or financial services, ideally with workers' compensation, underwriting, or claims workflows.
• Shipped systems that handle sensitive data…