Architecture Lead (AI Agent)

RedotPay; Hong Kong; On-site; HKD 900,000 - 1,500,000; Full time

Job summary

RedotPay is seeking an experienced AI Architect to lead the global customer-service AI Agent strategy and technical roadmap. You will design agent workflows combining LLMs with crypto payment flows, and build a RAG-based knowledge base to reduce hallucinations while ensuring compliance across the financial domain.

You will establish data-driven evaluation, define API/tool requirements, and collaborate with Hong Kong and global teams to enable multi-language support across 100+ regions.

Qualifications

• 3+ years in intelligent customer service / AI products or agent building in Fintech/Web3.
• Deep understanding of LLMs and their limitations.
• Hands-on with API design and tool integrations.
• Fluent in English and Chinese.

Responsibilities

• Lead overall planning and architecture design for RedotPay's global AI Agent and roadmap.
• Design multi-agent workflows and HITL hybrid architectures for support.
• Build RAG-based crypto knowledge base and guardrails for compliance.
• Set up data-driven eval framework and iterate with real-time feedback.
• Coordinate with global teams to roll out multi-language support across regions.

Education

Bachelor's degree or above

Tools

LangChain

LlamaIndex

Coze/Dify

Function Calling

APIs

Vector databases

Job description

RedotPay is a global crypto payment fintech integrating blockchain solutions into traditional banking and finance infrastructure. Our user-friendly crypto platform empowers millions globally to spend and send crypto assets, ensuring faster, more accessible, and inclusive financial services. RedotPay advances financial inclusion for the unbanked and supports crypto enthusiasts, driving the global adoption of secure and flexible crypto-powered financial solutions. Join us in shaping the future of finance and making a meaningful impact on a global scale.

Job Responsibilities

Agent Architecture Planning & Encrypted-Payment Technology Delivery

• Lead the overall planning and architecture design for RedotPay's global customer-service AI Agent (intelligent customer support / AI assistant), and develop a technical roadmap for intelligent customer service.
• Combine LLMs with Web3 / crypto payment business flows to design efficient Agent workflows.
• Design multi-agent collaboration mechanisms and human-in-the-loop (HITL) hybrid architectures for customer support.

Prompt Engineering & Web3 Knowledge Base (RAG) Building

• Precisely design and continuously fine-tune the system prompt and guardrails to ensure the Agent provides high-accuracy, compliant, and on-brand responses in the financial / crypto-assets domain.
• Lead the building of a RAG-based crypto payments knowledge base, integrating crypto knowledge, fiat on/off-ramp rules, and card scheme/compliance policies to effectively reduce LLM hallucinations.

Data-Driven Operations & Iteration for High-Frequency Customer Inquiries

• Set up a data-driven evaluation framework for global crypto users, including metrics such as:
• Use data tagging/analytics and log analysis to identify business pain points.
• Continuously iterate on the model, prompts, and knowledge base via real-time data and user feedback to improve the Agent's ability to handle complex long-tail issues (e.g., transaction delays, card declines/declines, on-chain status queries).
• Act as the bridge between business and technology, responsible for defining technical requirements for APIs / tool use / function calling
• Coordinate with Hong Kong and global teams to roll out multi-language AI customer service across 100+ countries/regions, including standard-setting.

Requirements

• Bachelor's degree or above (preferred in Computer Science, AI, FinTech, Data Science, etc.).
• 3+ years of hands-on experience in intelligent customer service / AI products / Agent building and operations in Fintech/Web3 or global payment fields
• Good to have deep understanding of core crypto payment processes and user pain points
• Deep understanding of LLMs (e.g., GPT-4, Claude, open-source models) including their limitations; proficient with Agent frameworks (e.g., LangChain, LlamaIndex, Coze/Dify), RAG architecture, and Function Calling / Tool Use solution design.
• Strong data analysis skills; experience with SQL / Python for data mining and model/prompt tuning is preferred.
• Familiar with API design, vector databases, and integrations/customizations with mainstream customer service ecosystems
• Fluent in English and Chinese
• High sensitivity and enthusiasm for Web3 / crypto-finance and AI frontier technologies.