Requirements & tools
Brex
The Software Engineer, Forward Deployed Agent Builder at Brex builds AI agents that automate and augment internal functions across the company. Responsibilities include embedding with partner teams to deeply understand workflows (including shadowing different job functions); scoping, designing, and deploying agents that take over real workflows across internal organizations; integrating agents with internal systems, APIs, and data sources; defining evaluation frameworks, success metrics, and feedback loops to measure and improve agent performance; and building shared tooling and playbooks that accelerate future deployments. The role works on top of existing models and frameworks, wiring together tools, MCPs, and internal systems into agentic workflows. Required: 4+ years in engineering with strong systems-thinking, hands-on experience shipping production AI/ML features, fluency in modern AI frameworks (LLMs, agent orchestration, RAG, MCP), and comfort embedding with non-technical operational teams. Brex serves customers including DoorDash, Coinbase, Robinhood, Zoom, and Reddit. Compensation $152,000-$240,000 base plus equity (RSUs), ESPP, generous health and 401(k) benefits.
Role context
Forward Deployed Engineers operate at the intersection of internal AI tooling and the operational teams whose workflows they automate. At Brex, this role embeds with finance, operations, and customer success teams to design, build, and ship AI agents that take over real workflows end-to-end. The position requires equal fluency in modern AI frameworks (LLMs, MCPs, agentic tools) and the operational reality of how work actually happens at a 200-market fintech platform. Compensation reflects Brex’s pay banding for technical IC roles at $152,000-$240,000 plus equity. Seattle-based three-day hybrid with up to four weeks of fully remote work per year.
Quick facts
Frequently Asked Questions
What does a Forward Deployed Engineer actually do day-to-day at Brex?
Day-to-day work splits roughly 40% embedded with partner teams (shadowing finance operators, customer success agents, AP clerks to understand their workflows in detail), 50% building (designing agent architectures, writing prompts, integrating internal systems, shipping production features), and 10% measurement (defining evaluation frameworks, monitoring agent performance, iterating on failure modes). The role is technical IC with high operational visibility.
What's the difference between Forward Deployed Engineers at Brex vs Palantir?
Palantir invented the Forward Deployed Engineer title for customer-embedded engineers who build custom integrations and analyses at client sites. Brex's version is internally focused — Forward Deployed engineers embed with Brex's own operational teams, not external customers. The skill profile overlaps (rapid building, operational empathy, systems integration) but the customer is internal, and the deliverables are reusable internal tools.
Is prior AI agent experience required?
Listed as required: 'fluency in modern AI frameworks (LLMs, agent orchestration, RAG, MCP)' and 'hands-on experience shipping production AI/ML features.' Engineers without AI agent specifics but with strong systems engineering backgrounds and demonstrable ability to ramp on new frameworks can be competitive — Brex's bar emphasizes builder mentality over specific framework expertise.