Software Engineer Intern: Agent Commerce

Root Inc.; United States; Remote; USD 30,000 - 52,000; Part time

Job summary

Root Inc. is seeking a software engineering intern to help build the Root Bind MCP server, enabling AI agents to quote, customize, and hand off auto insurance policies via our Bind API.

You will work on production-quality Python code, agent tooling, and discoverability for an emerging agent ecosystem, with a strong emphasis on AI coding assistants in daily workflows. The internship offers flexible, remote-friendly work across the US, with part-time hours around school and a full-time push during

Qualifications

• Currently pursuing a degree in Computer Science, Engineering, or related field.
• Comfortable writing code in Python or strong fundamentals and eager to learn Python.
• Familiarity with APIs, HTTP, and Git.
• Curious, self-directed with AI coding assistants.
• Strong written communication; work is documented and reviewed asynchronously.

Responsibilities

• Implement and test MCP tools in Python and FastMCP across the quote lifecycle.
• Integrate with Root's Bind API sandbox; handle tokens and async operations.
• Build an agent evaluation suite with transcripts for reliable quote flow.
• Improve server discoverability via MCP Registry and docs; write tool docs.
• Participate in code reviews and design discussions; present at internship showcase.

Skills

Python programming

APIs & HTTP

Git version control

AI coding assistants

Strong written communication

Education

Pursuing degree in Computer Science/Engineering

Tools

MCP (Model Context Protocol)

FastAPI

Docker

AWS

Job description

Root is on a mission to unbreak insurance by creating experiences people love at prices they can't believe. We believe that investing in world-class technology will facilitate a new class of insurance products, driving a massive positive impact on the hundreds of millions of drivers who carry auto insurance in the US.

The Opportunity

We're looking for a software engineering intern to help build the Root Bind MCP server: a Root-hosted server that lets consumer AI agents (like Meta's muse) quote, customize, and hand off auto insurance policies through our Bind API. You'll work alongside the AI Enablement team to ship real, production-quality code in Python, build agent tooling, and help make our service discoverable to the emerging agent ecosystem. This is a hands-on, AI-native role. We expect you to use AI coding assistants as a core part of how you work, you'll be using AI to help build the AI platform itself. Root is a “work where it works best” company. This means we will support you working in whatever location that works best for you across the US.

What You'll Do

• Implement and test MCP (Model Context Protocol) tools in Python and FastMCP that cover the quote lifecycle, coverages, legal disclosures, handoff, and post-handoff readback.
• Integrate with Root's Bind API against a sandbox environment, including token handling and async operations.
• Build and run an agent evaluation suite (golden-path transcripts) that proves an AI agent can drive the full quote flow reliably.
• Help make the server discoverable, MCP Registry, /.well-known/mcp.json, and llms.txt, and write tool documentation that agents can use without external docs.
• Participate in code review, design discussions, and team rituals, and present your work at the end-of-internship showcase.

What You’ll Learn

• How to design and ship production MCP servers and agent-facing APIs.
• How to evaluate and harden AI agent behavior against a real service.
• How a modern engineering team works — from design and code review to deployment and compliance.
• The auto insurance quoting and binding domain, and how Root keeps payment handling at PCI SAQ A.

What You’ll Need To Succed

• Currently pursuing a degree in Computer Science, Engineering, or a related field.
• Comfortable writing code in Python (or strong fundamentals in a similar language and eagerness to learn Python).
• Familiarity with APIs, HTTP, and version control (Git).
• Curious, self-directed, and comfortable working with AI coding assistants.
• Strong written communication, much of this work is documented and reviewed asynchronously.

Preferred Qualifications

• Experience with LLM APIs, agents, or tool-calling / function-calling.
• Familiarity with MCP, FastAPI, Docker, or cloud platforms (AWS).
• A personal or class project that shows you build things end-to-end.

Logistics

• Flexible, project-based arrangement: 550 hours over roughly three months.
• Part-time hours that fit around your school schedule to start.
• Full-time push during winter break (primarily December) to support the AI initiative.
• Reports to the Eng Strategy & Operations manager, with day-to-day technical mentorship from a Staff engineer on the AI Enablement team.

As part of Root's interview process, we kindly ask that all candidates be on camera for virtual interviews. This helps us create a more personal and engaging experience for both you and our interviewers. Being on camera is a standard requirement for our process and part of how we assess fit and communication style, so we do require it to move forward with any applicant's candidacy. If you have any concerns, feel free to let us know once you are contacted. We’re happy to talk it through.

Please see our Privacy Notice available HERE for more information on how we process your personal data.

Consistent with the Americans with Disabilities Act (ADA) and the Civil Rights Act of 1964, it is the policy of Root to provide reasonable accommodation when requested by a qualified applicant or candidate with a disability, unless such accommodation would cause an undue hardship for Root. The policy regarding requests for reasonable accommodation applies to all aspects of the hiring process. If reasonable accommodation is needed, please contact [e-mail].