Senior Corporate Engineering & AI Systems Engineer, Float

• You're a builder first — you've seen that the highest-leverage engineering right now multiplies everyone else, and you want to do it somewhere with real ownership and zero red tape
• 7+ years of software engineering, including internal platforms or tools that measurably increased team output
• 1–2+ years building LLM-powered systems in production — real users, real reliability requirements, not prototypes
• Strong production Python and/or TypeScript, plus the full-stack range to ship a usable interface when a workflow needs one
• Deep, practical knowledge of agentic systems: tool use / function calling, orchestration, context engineering, structured outputs, memory, and background-agent patterns
• An eval-first mindset — success metrics before you build; observability, guardrails, and cost controls as part of the system
• Integration and auth chops across an enterprise SaaS stack: REST, webhooks, event-driven patterns, OAuth / OIDC / SAML
• Security and governance judgment — scoped, auditable access, human-in-the-loop design, and a clear sense of what an agent should never do in a financial company
• Workflow-discovery skills: sit with a non-technical team and turn how they work into an automation spec with measurable targets
• Cloud fluency (AWS) and comfort with CI/CD

Desirable

• We're hiring a Senior Corporate Engineering & AI Systems Engineer to build the internal platform that lets every Floater scale themselves — so each person feels like a team of ten
• You'll own three things: the harness (the secure, governed layer connecting AI models to Float's internal systems, with identity-aware access, tool connectors, guardrails, and audit trails), the platform (the paved road that lets any Floater compose, run, and share automations), and the background agents (always-on workers that triage queues, reconcile data, and draft responses while Floaters sleep, escalating to a human when judgment is needed)
• This is not an evaluate-some-tools-and-write-a-policy role — you'll be in the room when a team describes a problem, and shipping the fix, often the same week
• Build the harness: authenticated, scoped, auditable connectors and MCP servers linking LLMs to Slack, Google Workspace, Salesforce, NetSuite, Zendesk, and Float's own product
• Ship background agents that run multi-step workflows end to end — with the tool routing, memory, retries, human-in-the-loop approvals, and audit trails that make them trustworthy in fintech
• Build the skills-sharing layer that turns one team's automation into everyone's capability, and drive the adoption that makes it worth building
• Embed with teams to map how they work and turn ambiguous problems into shipped systems — defining evals before you build and instrumenting reliability and cost as you go
• Own governance with Security and Risk: non-human identity, scoped permissions, prompt-injection defense, and privacy guardrails that enable rather than block