01
Durable control-plane services
Build and refine the API, scheduler, projectors, workers, runner, Slack ingress, and migration paths that keep agent work recoverable.
Careers / Engineering
Build the control plane for useful agents.
This is the first developer role on the z0 codebase: a backend-heavy full-stack engineer who can make agent work durable, inspectable, and safe enough for real teams.
Stage
Pre-seed
Location
NYC
Level
Senior
Comp
$140–180K + 1.0–2.5%
Codebase
Go 1.26PostgreSQLNext.js 15React 19VitePulumi
Platform surface
Infrastructure with a product face.
z0 sits between people asking for work in Slack, agents doing the work, and operators who need scope, cost, approval, and evidence to stay clear.
Scope
Backend-heavy, full-stack product engineering across six surfaces with one operator bar.
02
PostgreSQL event flows
Evolve the event store, outbox claiming, read models, usage ledger, and tenant-scoped records without losing auditability.
03
Policy and approvals
Make tool access, network destinations, budget checks, denials, and human review clear before an agent reaches sensitive systems.
04
Agent runtime execution
Improve Codex and Claude Code adapters, prompt packaging, checkpoint recovery, sandbox snapshots, and retry behavior.
05
Slack and integrations
Keep Slack routes, Composio-backed tools, tenant installs, approved destinations, and customer-owned credentials on controlled rails.
06
Operator-facing surfaces
Ship precise React interfaces for onboarding, workspaces, agent setup, usage, approvals, and failure inspection.
Open roleAgent control plane
Founding Full Stack Engineer
z0 is building the operating layer for teams that want AI agents to do real work from Slack without giving up control. The product combines tenant and workspace scope, Slack routing, approved integrations, memory, budget admission, human review, sandboxed execution, and durable audit records.
This role is for a developer who can own hard paths across the stack. You will work in Go services, PostgreSQL migrations, OpenAPI contracts, runner adapters, telemetry, and React UI where admins and operators inspect what happened.
Responsibilities
- Ship production changes across Go services, database migrations, OpenAPI contracts, and React product surfaces.
- Design event-driven workflows where retries, interruptions, approvals, and settlements are explicit records.
- Harden tenant and workspace authorization so Slack routes, memory, tools, and usage never cross scope boundaries.
- Improve runner reliability through checkpointing, sandbox lifecycle work, deterministic local execution, and strong telemetry.
- Turn operational questions into inspectable product flows for admins, approvers, auditors, and support staff.
- Write tests at the level of risk: unit coverage for policy and accounting, integration coverage for orchestration paths, and UI coverage for customer flows.
You might be a fit if
- You have taste. It shows up in the interfaces you build and in code that reads like someone knew what to leave out.
- You have shipped Go or TypeScript systems that keep working when jobs retry, processes crash, or webhooks arrive twice.
- You are comfortable with PostgreSQL schema design, transactions, idempotency, queues, event logs, and read-model projection.
- You can move from backend invariants to React UI details without treating either side as someone else's problem.
- You care about security boundaries, audit trails, least privilege, and clear operator evidence.
- You prefer direct, testable code over fashionable abstractions.
Apply
Bring judgment,
not ceremony.
Send a short note with the systems you have shipped, the failures you have debugged, and the parts of this codebase you would want to own first.
join@z0.ai · replies within a few days