Integrations Engineer

About HASH

At HASH, we're building infrastructure to solve information failure and help everybody make the right decisions. Our open-source platform helps organizations turn fragmented information into typed, temporal knowledge and process graphs that support analysis, simulation, optimization and automation.

We're building towards world models from a different starting point from most AI labs: explicit representations of real systems, the entities within them and the processes through which they change. Our founders include Joel Spolsky, who founded Stack Overflow and Trello;
Jude Allred, formerly CTO of Glitch; and Dei Vilkinsons, formerly CEO of Soho Strategy.

A large part of HASH's utility is its ability to integrate structured data from a wide-range of external applications into a single, unified knowledge and process graph. Each SaaS applications, communication tools, internal service or specialized enterprise system has its own authentication model, API conventions, rate limits, schemas, failure modes and ideas about time. HASH needs to connect to these systems reliably, preserve where information came from and keep its model synchronized as the source changes.

We're hiring an Integrations Engineer to work on these problems.

You will design and implement connectors in Type Script and Python, making both widely used services and customer-specific systems straightforward to connect to HASH. This includes understanding each source in detail, retrieving and synchronizing its data reliably, and mapping its native concepts into common semantic schemas that remain useful beyond a single integration. While some services provide excellent public APIs, complete schemas and thoughtful documentation...

others provide only incomplete documentation, inconsistent endpoints or no public documentation for the behavior we are authorized to integrate with. You should enjoy both situations. You may find yourself studying an SDK, inspecting browser network traffic, comparing payloads, testing undocumented behavior and working out how an application really functions—then turning that understanding into a connector that is robust, respectful of permissions and maintainable as the service changes.

Different systems use different names, structures and assumptions for the same real-world concepts, while superficially similar fields may mean different things. You should find the idea of how a Salesforce account, an SAP business partner, a row in a customer database and an organization found on the web can map into shared types without erasing their provenance or source-specific meaning.

This is not a role for writing disposable scripts, and you will be supported in the role by a team working on the surrounding integration infrastructure. Your main output will be excellent connectors: new ones developed quickly and carefully, and existing ones kept working as upstream services and HASH evolve.

You will work closely with core platform engineers, AI Success Engineers and customer-facing colleagues. Sometimes you will be adding support for a widely used SaaS product. Sometimes you will need to understand a customer's unusual internal API or data model. In both cases, your responsibility is to produce something accurate, maintainable, secure and observable enough to remain dependable after the initial implementation.

You will also be expected to leverage AI effectively in this role.

• Strong production experience with both Type Script and Python
• Experience personally building and maintaining multiple connectors to third-party APIs, SDKs, databases or services
• A solid understanding of HTTP, REST, GraphQL, webhooks, OAuth 2.0 and common authentication patterns
• A genuine appetite for reverse-engineering poorly documented integration surfaces by reading client code, inspecting network behavior, testing hypotheses and comparing real payloads
• Experience handling the realities of external systems: pagination, rate limits, partial failure, retries, deduplication, idempotency, eventual consistency and breaking API changes
• An obsession with semantic modeling: the ability to identify the real-world concepts behind unfamiliar data and map them into common types without discarding source-specific meaning
• Experience building asynchronous or event-driven services, including queues, workers, scheduling and backfills
• Strong testing instincts, including unit, integration, contract and end-to-end testing where appropriate
• The ability to diagnose failures across several systems using logs, metrics, traces and careful reproduction
• Good security judgment around credentials, secrets, permissions, tenant isolation and sensitive customer data
• High agency and the ability to take an integration from incomplete documentation and a real customer need to a dependable production implementation
• Clear written and spoken communication, including the ability to work directly with customers and external technical teams when necessary

You do not need experience…