Senior Software Engineer, Control Plane

We're hiring a senior full-stack engineer for a product where the data layer is the product: a dozen AI vendors with a dozen billing models, the systems engineering teams run on, device fleets and identity directories, all reconciled into one model that a CFO signs and an engineering leader acts on. Next.js and Postgres are the tools; the job is the domain.

About ZuzAI

AI coding agents and assistants are changing how software is built, and they are becoming one of the fastest-growing, least-understood lines on an enterprise's bill. Engineering leaders can't yet answer the basic questions: what it costs, what it produces, and whether it's being used the way the company intended. ZuzAI builds the infrastructure that answers them. We're deployed in production at global technology companies, inside their own clouds, and we ship to them every week.

We're a small, senior, agent-native engineering team, on-site (3 days a week) , funded, and early enough that what you build defines the product. You'd work directly with the Director of Engineering and have a direct line to the founders.

01 The challenge

Large engineering organisations are adopting AI coding agents and assistants faster than they can understand what they cost, what they produce, and who is allowed to do what. ZuzAI builds the infrastructure that gives them control. The control plane is where that control lives, and it is hard in ways a typical web app isn't:

• The data layer. Events from AI providers, source control, issue trackers, delivery pipelines, incident tools, HR and identity directories, device management, and our own pipeline, each with its own identifiers, time zones and ideas of a "user", unified into one taxonomy of people, units, tools, models, sessions and work that every number is built on.
• Ingestion at scale. Sync engines against rate-limited vendor APIs, webhooks, backfills, late data and idempotent retries; rollups and partitions that keep the console fast at the scale of thousands of engineers.
• AI tokenomics. Input, output and cached tokens, credits, overage, pools, seat tiers, model price cards, committed spend, and the gap between what a vendor reports and what it invoices. You will know how every AI vendor bills, and reconcile usage to the cent.
• ROI of AI spend. Connecting what was spent on coding agents to what engineering shipped: attribution rules that hold up under a sceptical VP, across systems that never agreed on an identifier.
• Enterprise-ready. SSO (OIDC and SAML), SCIM and lifecycle, RBAC and scoped access, an audit log with export to the customer's SIEM, OpenTelemetry export, retention and residency, encryption and key management; deployed inside the customer's own cloud.
• Performance. Analytical Postgres under load, query plans you can read, caching and memoisation with honest staleness, a console that stays fast as data grows.
• The business domain is yours. FinOps, engineering intelligence, vendor billing: you'll talk to customers' finance and engineering leaders and define what each number means before you build it.

02 Why it's interesting

• Greenfield modules with real users from day one: design the model, build the engines, ship the page, watch an executive use it the same month.
• A feedback loop measured in days: we ship to enterprise production weekly.
• Problems with no textbook answer, in a domain that is being defined right now.
• A small, senior, agent-native team, working directly with the CTO.

03 What you'll do

• Own modules end to end: the taxonomy and schema, ingestion and sync engines, rollups, APIs and the pages people use every day.
• Build and keep honest the integrations: AI vendors' admin and billing APIs, source control, trackers, delivery, incidents, identity and device sources.
• Make the money right: pricing models per vendor, reconciliation to invoices, attribution to people, units, repos and models.
• Ship the enterprise features: SSO/SAML, SCIM, audit and SIEM export, OpenTelemetry, retention, with the security posture a CISO review expects.
• Hold the bar: strict tenant and role isolation, end-to-end tests on every workflow, performance budgets, reviews that make the team faster.

04 What we're looking for

• Years of building and owning data-heavy products in production, from schema to screen; you've designed a data model that other people's numbers depended on.
• Strong TypeScript, React, Next.js and Node.js, with taste for dense, fast, readable product UI.
• Deep Postgres: data modelling, rollups, partitioning, indexing, query tuning over millions of rows; Prisma and raw SQL both.
• You've built ingestion pipelines and third-party integrations and know where they break: rate limits, pagination, backfills, idempotency, late data.
• Curiosity for the domain: billing models, token economics, model pricing, how engineering organisations measure themselves.
• You've shipped enterprise security features before, or want to: SSO, RBAC, audit, export to customers' systems.
• Product judgment: you take a rough problem from the CTO or a customer, work out the model, the edge cases and the page, and ship the whole thing.
• You build with coding agents as your normal way of working, and you review their output as carefully as your own.

05 Nice to have

• FinOps, engineering-intelligence, observability or billing products.
• OpenTelemetry, SIEM or data-export integrations; device management (MDM) data.
• AWS, Docker, Terraform, and products deployed in customers' own clouds.
• You've been early at a company before and liked it.