Senior Software Engineer, AI Agents

About ReapReap is a global financial technology company headquartered in Hong Kong with employees across multiple countries. We enable financial connectivity and access for businesses worldwide by combining traditional finance with stablecoins for efficient money movement.

Through our stablecoin-powered corporate cards, payments, and expense management tools, we streamline financial operations and help businesses scale. Our APIs enable businesses to integrate stablecoin-enabled finance into their own products and services—from issuing Visa cards to facilitating cross-border payments.

Backed by leading investors including Acorn Pacific, Index Ventures and HashKey Capital, Reap is building the future of borderless, stablecoin-enabled finance.

Why ReapWe're building the AI CFO — an intelligent, proactive agent that helps businesses manage their money before problems arise, not after. We sit on rich financial data and real operational workflows. Now we're putting AI at the centre of that.

This isn't starting from zero. Our card operations agent is about to go live internally — we dogfood our own product, so you'll see the impact of what you build firsthand before it reaches external clients. Your job is to take it from 0.1 to 1 — refining reliability, building observability into every layer, and making the decisions that turn a working prototype into a production-grade system businesses trust with their financial operations.

You'll work alongside our AI Product Manager to expand from card operations into accounting automation, policy enablement, and beyond — tackling the workflows that drain finance and finops teams of time and energy. Every agent you build needs to be compliant, reliable, and handle PII with the seriousness that fintech demands. You'll also need to think critically about unit economics and adoption — an agent nobody uses or that costs more than it saves isn't worth shipping.

If you come from a fintech background and understand why PII handling isn't optional, that's ideal. If you don't but you have the chutzpah to figure it out and get genuinely excited every time a new model drops, we want to talk to you too.

Your Mission Awaits

• Take our card operations agent from internal pilot to production — building the reliability, observability, and guardrails needed for a system handling real financial data and PII.

• Own the tradeoffs between latency, model selection, cost, and safety — making pragmatic architectural decisions that keep our unit economics viable as we scale.

• Build agent systems that are proactive, not reactive — designing solutions that anticipate what finance teams need rather than waiting to be asked.

• Expand agent capabilities across accounting automation, policy enablement, and card operations — working with the AI Product Lead to prioritise what moves the needle for adoption.

• Stay sharp on the frontier of LLM research and tooling — evaluate new models, methods, and architectures and bring what works into our stack.

Think like a product engineer, not just an AI engineer — every system you build should drive platform adoption and make clients' lives measurably easier.

Your Superpowers

• 8+ years of experience in full-stack or backend development — with strong proficiency in Python as your primary stack. Experience with Java, Golang, or Rust is equally welcome.

• Strong software engineering foundation — experience designing distributed systems, APIs, and scalable backend architectures.

• 1–2+ years of hands-on experience building AI agents in production — you've gone beyond prompting LLMs and have shipped systems that reason, plan, and act.

• Deep understanding of LLM internals — you know how models work under the hood, not just how to call an API.

• Architectural thinking beyond frameworks — you can evaluate when LangChain, LangGraph, AutoGen, or other agent frameworks are the right tool, and when to build from first principles.

• Familiarity with Model Context Protocol (MCP) — you understand what MCPs are and how they enable agent-tool interoperability.

• Leadership and delegation instincts — you're comfortable guiding and reviewing the work of others, and you operate with ownership over outcomes, not just tasks.

Excellent communication skills — you can translate complex technical decisions into clear reasoning for engineering and product stakeholders.

Nice to Have

• Experience in fintech, payments, or financial services (though strong engineers adapt — domain knowledge can be learned).

• Proficiency in Node.js and TypeScript.

• Familiarity with MLOps practices and AI deployment pipelines.

• Experience with multi-agent coordination, tool use, and memory architectures.

Prior exposure to agentic evaluation and observability tooling.

Why You'll Love it Here

• A high-impact role in a rapidly growing fintech startup

• Flexible hybrid work environment with a global, collaborative team

• Insurance coverage after probation

• Reap Card stipend

• Use of AI