Harper

Operating Memory Lead

San Francisco · Exec

Sponsorship not specified$90k-$150kDetected 55 days ago

RESTLogisticsResearchCommunicationWritingUnderwriting

About the role

The bet - turning human judgment into compute - has a precondition: the judgment has to be written down.
AI doesn't magically understand a company.
It works only when the business is documented clearly enough for systems to retrieve the right context, recognize the workflow, handle the edge cases, and escalate when a human is actually needed.

Capture tribal knowledge. Embed with sales, intake, service, placements, and renewals. Sit with operators, shadow workflows, listen to calls, read transcripts, and document what people "just know."
Build operating memory. Turn transcripts, Slack threads, Looms, and one-off explanations into source-of-truth docs, decision logs, playbooks, process maps, onboarding paths, and glossaries.
Use AI as a force multiplier. Build repeatable workflows that turn raw context into decisions, owners, open loops, SOPs, training material, and product requirements.
Ask the questions that turn a meeting into an artifact: who owns this, what's the exception, what's the source of truth, what would someone misread from the transcript alone.
Find the edge cases. Document where workflows break - reworks, escalations, stale quotes, underwriter follow-ups, payment/binder gaps, COI delays, customer confusion.
Maintain the knowledge base. Keep docs current, assign owners, kill stale guidance, make sure people know where the truth lives.
Sit with operators, shadow workflows, listen to calls, read transcripts, and document what people "just know."
Almost no one joins Harper for insurance; they join to build the company that replaces how it works.

demonstrated ability to interview stakeholders and extract operational detail
comfort in a fast-moving, ambiguous startup
Genuinely AI-native in practice - not "I use ChatGPT," but you have taste for when an output is structurally wrong, not just stylistically off.
You prompt for extraction (decisions, contradictions, owners, edge cases), not just summarization, and you know good output depends on good context you engineer upstream.

Uber commuter benefits; breakfast, lunch, and dinner provided; snacks, drinks, and coffee daily; free gym membership; health, dental, and vision insurance.
Bonus points if you include the artifact - doc, playbook, process map, onboarding guide, research synthesis, curriculum, or internal system - and show how you used AI tools to do it faster or better.

how a top rep prioritizes quotes, how service handles an edge case, which underwriter to chase, what a customer really means when they push back at bind, why a workflow changed yesterday.
That works at small scale and breaks at ~1,000 new customers a month.
The next bottleneck here isn't engineering - it's knowledge.
Every process that lives only in someone's head is a future failure mode.
Every undocumented edge case is rework.
A workflow that isn't clear enough for a new hire isn't clear enough for an AI agent either.
You turn that messy operating reality into structured, AI-legible knowledge - and make sure Harper's knowledge compounds instead of disappearing.
Real AI-tool fluency is a hard gate, not a nice-to-have. You'll be judged on whether your AI-built artifacts are structurally correct and reusable - taste matters more than tool familiarity.
You succeed only if behavior changes. A beautiful doc nobody uses is a failure here. The measure is whether the workflow got faster, the edge case stopped recurring, the new hire ramped without a meeting.
On-site SF, long days. Mon-Fri, in-office hours that match the rest of the company.
Monday-Friday, in-office hours matching the rest of the company.
We're not bolting AI onto insurance - we're rebuilding the entire business as software, on a simple bet: turning expert human judgment into compute is one of the largest transitions left to make, and a trillion-dollar industry still run 90% by hand is the place to prove it.

This listing is sourced directly from Harper's careers page and normalized into a canonical job model.