Staff Software Engineer (Remote)

About the role

FirstWork is building the next generation of agentic AI products for frontline hiring, the work that gets roughly 80% of the world's workforce hired, verified, and paid. We're looking for a Staff Backend Engineer to architect the systems that let those agents run at the next order of magnitude, and to build the new product lines we are launching with AI designed-in from the start.

In this role, you'll set the technical direction for a core domain and work closely with product, design, and engineering leadership to get intelligent systems into a shape customers can configure and rely on. This is a hands-on architecture role: you'll still write code, and you'll see your decisions tested against live enterprise volume. The challenges ahead are scaling throughput, latency, and data volume as usage compounds, and keeping AI agents accurate, reliable, and economical as that happens.

The scale we run at

-

~50,000 workers a week move through our hiring and onboarding flows.

-

2.5 million API requests a day across the platform.

-

Millions of document pages a year processed by AI agents, across 40+ countries and a dozen-plus languages.

-

Global staffing, logistics, healthcare, marketplace, and employment platforms across the US, UK, EU, LATAM, and India.

-

Billing on work performed, per document validated and per agent action executed, which makes product design, cost per unit, and gross margin one conversation.

What you'll do

-

Take end-to-end ownership of core domain: the agent runtime, the compliance and rules engine, the worker data model, or integrations.

-

Push LLMs and agentic AI deeper into the product, and draw the line between what belongs in inference and what belongs in deterministic, auditable rules.

-

Design the safety properties of agents that act in external systems, including idempotency, retries, isolation, and human escalation.

-

Build the evaluation and accuracy infrastructure that makes AI decisions defensible in a regulated context.

-

Keep inference viable at scale through model routing, caching, and batching, with cost per transaction as a tracked metric.

-

Design data models with long-lived constraints, including audit logs, compliance clocks, and credential lifecycles.

-

Drive the data architecture roadmap: read replicas, logical sharding, CDC-fed projections, and a search tier.

-

Enforce multi-tenancy, permissions, and audit trails as invariants guaranteed by design and by test.

-

Act as a technical lead through design review, incident command, and mentoring senior engineers.

-

Work directly with product and customers to turn bespoke requests into platform capabilities.