Member of Technical Staff - NYC

Corridor; New York (NY); On-site; USD 150,000 - 230,000; Full time

Job summary

Corridor is seeking founding AI engineers to shape the architecture, product, and engineering culture of an AI-native benefits platform. You will own end-to-end problems from data models to end-user experiences, in a high-impact domain where getting things right matters for real people.

Strong proficiency in Python and TypeScript is expected; you’ll work on frontier agentic systems and tools that scale to millions of users, with an emphasis on reliability, guardrails, and safety.

Qualifications

• Strong proficiency in Python and TypeScript.
• Experience building AI systems with reliability and safety considerations.

Responsibilities

• Own end-to-end problems from data models to end-user experiences.
• Build frontier agentic systems and tools to scale to millions of users.
• Ensure reliability and safety with guardrails and production-ready behavior.

Skills

Python

TypeScript

Job description

About Corridor

Corridor is the first AI-native employee benefits brokerage. By leveraging an internal harness of AI agents, we tackle some of the hardest problems in healthcare to the benefit of our users. This is a true AI-native services business, backed by blue-chip venture funds.

Our mission is to make health insurance and benefits actually work for the people who navigate them — the operators who place coverage and the employees who live with it. It’s one of the largest, most consequential, and most poorly served markets in the country, and the tools for it haven’t meaningfully changed in decades.

We’re building SOBA, an AI-powered benefits platform, and the agentic systems at its core. Our team is small and talent-dense. We’re looking for founding AI engineers to help shape the architecture, the product, and the engineering culture, with the ownership and equity that implies.

The role

Engineers at Corridor are not feature builders. You’ll work on some of the hardest open problems in applied AI: how do you build agents that run a full broker analysis across messy, real-world inputs — where there are many valid paths and no single right answer — reliably enough that an operator, and the families who depend on the outcome, can trust the result? You’ll own problems end to end, from the data model to the end-user experience, on systems where being wrong has real consequences for real people.

What you’ll build

• Frontier agentic systems that execute broker analyses and operational tasks in high-degree-of-freedom environments under strict reliability requirements.

• Agents and tools that let operators serve orders of magnitude more clients while delivering higher-quality service to each — quality enforced through guardrails and guarantees, not hope.

• Applications that give individual employees unprecedented access to the information they need to choose and use their benefits well.

• Orchestration and infrastructure that coordinates agents across complex workflows and scales reliably to millions of users.

Exceptional candidates have demonstrated

• Technical ownership. You see problems through from ideation to impact — not just shipping a feature, but making sure it moves the thing it was meant to move.

• Systems judgment for high-stakes reliability. Strong opinions about correctness, failure modes, and production behavior in systems where the inputs are messy and the stakes are high.

• Good calls in the gray area. You can weigh data, user experience, and product taste when there’s no single right answer — and make progress on hard problems with incomplete specs.

• Willingness to get in the weeds. The most valuable problems here live in the details of how insurance actually works and how operators and members actually behave. You want to learn that.

• Curiosity about agents and AI. You’ve dug into how LLMs work, how agents fail, and what it takes to constrain them so they’re safe to put in front of users.

• Velocity without shortcuts. A track record of shipping fast while holding the code quality a high-density team expects.

• Mission and impact. You care that the work changes people’s access to care and benefits — not only that it’s technically interesting.

Strong proficiency in Python and TypeScript (our primary languages, with React and MCP) is what we work in day to day, but we care more about how you think and what you’ve shipped than about a checklist.