Back Office Engineer

Location In-person positions are available in San Francisco, or remote opportunities are offered for candidates in Canada. We do not hire remote employees in the US, only in Canada. The Role Truss runs a custom-built back office that powers our reviews, case work, and risk decisions. It works — but most of it still runs through human review, and the volume is only going up. We're hiring a Back Office Engineer to change that. Your job is to build the automation layer on top of our back office. Most of the reviews and cases that run through it today shouldn't need a human. Your work will: automate the routine reviews end-to-end, build the systems that let us handle cases at the scale we're growing into, and design the architecture that catches fraud before it happens rather than after. This is a high-leverage applied AI role. The systems you build will run at high volume against real operational workloads, so quality, observability, and the right guardrails matter from day one. We want someone who can move fast, build things that hold up under scrutiny, and treat the back office as a serious engineering surface — not a backwater. You'll work directly with the founder. Truss is a small team — 11 people today — so there's no layer between you and the decisions. What you ship matters, and you'll see the impact immediately. What You'll Do Automate back office reviews. Take the high-volume, repetitive review work happening in our custom back office today and replace it with reliable, auditable automation — with clear human-in-the-loop escalation paths for the cases that need judgment. Build automation to handle cases at scale. Design the workflows, queues, and decisioning logic that let case volume grow without growing the ops team linearly. This means thinking carefully about routing, prioritization, SLAs, retries, and audit trails. Build the architecture to catch fraud before it happens. Move our posture from reactive review to proactive prevention — instrumentation, signal aggregation, and real-time decisioning that stops bad activity before it lands on someone's queue. Apply AI thoughtfully. LLMs and other models are the right tool for a lot of this work, but they need evaluation harnesses, guardrails, and clear behavior when they're uncertain. You'll own that whole loop, not just the prompt. What We're Looking For Minimum 3 Years Of Professional Software Engineering Experience. Strong instincts for automation and workflow systems — queues, state machines, retries, idempotency, audit logging. You know how to build software that holds up under scrutiny. Hands‑on experience shipping AI/LLM-driven features in production, with real opinions about evaluation, guardrails, and when AI is and isn’t the right call. A pragmatic, ops‑aware working style. You enjoy working alongside non‑engineers, can translate fuzzy operational pain into shippable software, and care about whether the thing you built actually saved someone time. Payments or fintech experience is required. You should have shipped production software in a payments, fintech, or adjacent regulated financial environment, and be comfortable with the domain — chargebacks, KYC/AML, dispute lifecycles, fraud detection, or similar. Nice to Have Experience building fraud prevention or risk decisioning systems. Background working closely with operations, trust & safety, or compliance teams. Prior experience with evaluation pipelines for LLM-based decisions in regulated environments. #J-18808-Ljbffr Bonfirevc