Senior AI Engineer; Document Extraction

About hyperexponential (hx)

At hyperexponential, we’re building the AI‑powered platform that enables the world’s most critical decisions in a $7 trillion industry, which risks to take, and how to price them. These are the decisions that shape real‑world outcomes: whether rockets successfully launch into space, autonomous vehicles make it to market, or communities recover after major storms.

Until now, insurance has been making billion‑dollar decisions using outdated tools. We’re changing that. Our platform brings together data, AI, and human expertise to give insurers the fastest path from submission to decision - helping them move faster, act smarter, and take on more risk with confidence.

Backed by a16z, Highland Europe, and Battery Ventures, we’re scaling globally - already trusted by nearly 50 of the world’s largest insurers, with zero churn and billions in premiums flowing through hx.

What began as a single product in one market has rapidly evolved into a multi‑product, multi‑territory platform powering every stage of pricing and underwriting. AI is at the core of what we do - from building the world’s first domain‑specific AI peer programmer for insurance (think Git Hub Copilot with a PhD in actuarial science) to shaping agentic workflows that reinvent how this industry operates.

What makes hx different is the people who build it. Here, impact isn’t tied to title or tenure; it’s defined by the challenges you take on and the discipline you bring. Surrounded by peers who stretch you, you’ll do the best, hardest work of your life in a company engineered to endure.

If that sounds like you, join us in building what comes next.

AI is central to hx’s ambition to reinvent how insurers move from submission to decision. A significant part of that challenge begins before pricing even starts: transforming messy, inconsistent, insurance‑specific documents into structured, reliable data that underwriters can trust. The AI Ingestion team owns this critical first mile.

Within this team, we build and scale document extraction and data transformation systems tailored to the complexity of specialty insurance. We operate at the intersection of large language models, classical document understanding techniques, evaluation‑driven design, and production‑grade infrastructure. Our outputs directly determine whether insurers can automate workflows across multiple lines of business with confidence.

As a Senior AI Engineer focused on Document Extraction
, you will take ownership of solving high‑stakes, insurance‑specific extraction problems. Over the next 6 to 12 months, your mission will be to automate key data extraction and cleansing workflows across core specialty lines, hitting strict accuracy and consistency benchmarks while ensuring our systems are robust, observable, and built to scale.

• Designed and delivered high‑accuracy document extraction pipelines that automated key data capture across multiple specialty lines of business, enabling measurable reductions in manual processing during customer go‑live.

• Built and productionised data cleansing and manipulation capabilities within our ingestion system, improving consistency and reliability of downstream pricing inputs.

• Defined and enforced rigorous evaluation‑driven development practices, creating robust evaluation sets and benchmarks that guided model and pipeline improvements.

• Improved our evaluation infrastructure for complex ingestion scenarios, enabling faster iteration cycles and clearer visibility into performance trade‑offs.

• Partnered directly with customers during go‑live, diagnosing and resolving integration bottlenecks and production issues to ensure smooth technical rollouts.

• Led technical deep‑dives and brainstorming sessions to unblock complex extraction challenges, raising the bar on experimentation, knowledge sharing, and solution design across the team.

• Built and deployed document extraction or document understanding systems that achieved measurable accuracy improvements in production environments.

• Designed evaluation frameworks and datasets that enabled rigorous, repeatable…