About Harbor
Harbor is insurance for the agentic economy. Banks, insurers, hospitals and law firms are handing real decisions to AI agents: approving loans, pricing policies, moving money, advising customers, thousands of times a day. When one gets it wrong, nobody knows who is liable, and insurers are excluding AI from existing policies because there is no claims history to price it.
 
We fix that. Our platform finds every agent a company runs, measures its operational, security and regulatory risk against a framework mapped to NIST AI RMF, SR 26-2 and the EU AI Act, and blocks the decisions that break the company’s rules before they go through. Then we insure the risk that remains, with our own money behind the answer.
 
We emerged from stealth in October 2026 with $10M in seed funding from Base10 Partners, Abstract and 8VC. Born out of Stanford, Harbor brings together AI researchers, engineers, IMO and IOI medalists, former founders, and insurance leaders, with the shared mission of building the risk infrastructure for AI.
 
About the role
We are hiring Members of Technical Staff across research, product and agents. We use one title because the work is one problem: making AI agents safe enough to insure. You might spend a quarter building the evaluation harness that decides whether a bank’s agent gets covered, and the next one shipping the runtime that blocks a bad decision before it reaches a customer.
 
What you will work on

• Agent discovery: finding every agent a company runs, across clouds, vendors and shadow deployments.

• Measurement: evaluations, red-teaming and risk scoring against our framework, with results that an underwriter can price and a regulator can read.

• Mitigation: the runtime controls that stop an agent breaking a rule, at machine speed, without breaking the agent.

• The platform itself: the product our customers log into every day, the SDK and integrations their engineers use, and the data that connects all of it.

Who we are looking for

• You have built real systems that real people depended on, and you can go three levels deep into how they worked and how they failed.

• Strong in at least one of: ML and LLM systems, distributed systems and infrastructure, security, or product engineering with taste.

• You have shipped software to production and carried a pager for it.

• You are comfortable in a codebase where the spec is a conversation with a customer and the deadline is their launch date.

• You want to work on agents from the angle nobody else is taking: not making them smarter, making them insurable.

Nice to have

• Experience with agent frameworks, tool use, evaluations or LLM red-teaming.

• Background in regulated industries, risk, or security.

• Competitive programming, research publications, or a company you started.

How we work

• Founder mentality: Everyone owns a piece of the company, not a job description.

• Ownership over outcomes: You decide, you ship, you answer for the result.

• High pace: We decide in days, ship in weeks, and fix what we got wrong just as quickly.

• Good hearts: Direct, demanding and kind.

• Hybrid in San Francisco.

What we offer

• Competitive salary and meaningful early-stage equity.

• Medical, dental and vision for you and your dependents.

• Unlimited paid time off.

• Relocation support to San Francisco.

Harbor is an equal opportunity employer. We hire on what you can do. If you need an accommodation during the process, tell us and we will arrange it.

About Harbor
Harbor is insurance for the agentic economy. Banks, insurers, hospitals and law firms are handing real decisions to AI agents: approving loans, pricing policies, moving money, advising customers, thousands of times a day. When one gets it wrong, nobody knows who is liable, and insurers are excluding AI from existing policies because there is no claims history to price it.
 
We fix that. Our platform finds every agent a company runs, measures its operational, security and regulatory risk against a framework mapped to NIST AI RMF, SR 26-2 and the EU AI Act, and blocks the decisions that break the company’s rules before they go through. Then we insure the risk that remains, with our own money behind the answer.
 
We emerged from stealth in October 2026 with $10M in seed funding from Base10 Partners, Abstract and 8VC. Born out of Stanford, Harbor brings together AI researchers, engineers, IMO and IOI medalists, former founders, and insurance leaders, with the shared mission of building the risk infrastructure for AI.
 
About the role
We are hiring Members of Technical Staff across research, product and agents. We use one title because the work is one problem: making AI agents safe enough to insure. You might spend a quarter building the evaluation harness that decides whether a bank’s agent gets covered, and the next one shipping the runtime that blocks a bad decision before it reaches a customer.
 
What you will work on

• Agent discovery: finding every agent a company runs, across clouds, vendors and shadow deployments.

• Measurement: evaluations, red-teaming and risk scoring against our framework, with results that an underwriter can price and a regulator can read.

• Mitigation: the runtime controls that stop an agent breaking a rule, at machine speed, without breaking the agent.

• The platform itself: the product our customers log into every day, the SDK and integrations their engineers use, and the data that connects all of it.

Who we are looking for

• You have built real systems that real people depended on, and you can go three levels deep into how they worked and how they failed.

• Strong in at least one of: ML and LLM systems, distributed systems and infrastructure, security, or product engineering with taste.

• You have shipped software to production and carried a pager for it.

• You are comfortable in a codebase where the spec is a conversation with a customer and the deadline is their launch date.

• You want to work on agents from the angle nobody else is taking: not making them smarter, making them insurable.

Nice to have

• Experience with agent frameworks, tool use, evaluations or LLM red-teaming.

• Background in regulated industries, risk, or security.

• Competitive programming, research publications, or a company you started.

How we work

• Founder mentality: Everyone owns a piece of the company, not a job description.

• Ownership over outcomes: You decide, you ship, you answer for the result.

• High pace: We decide in days, ship in weeks, and fix what we got wrong just as quickly.

• Good hearts: Direct, demanding and kind.

• Hybrid in San Francisco.

What we offer

• Competitive salary and meaningful early-stage equity.

• Medical, dental and vision for you and your dependents.

• Unlimited paid time off.

• Relocation support to San Francisco.

Harbor is an equal opportunity employer. We hire on what you can do. If you need an accommodation during the process, tell us and we will arrange it.