Could AI Engineer

Role overview

A specialist engineering role focused on building AI-native tools for legal and compliance teams inside a large global cryptocurrency and blockchain platform. The position partners closely with Legal, Compliance, and the Money Laundering Reporting Officer to design retrieval, agentic, and evaluation systems that make regulatory workflows faster and more reliable. It suits a hybrid engineer who understands both compliance operations and modern LLM architecture, and who can prototype quickly in ambiguous problem spaces.

Responsibilities

- Work with Legal, Compliance, and the MLRO to support internal compliance policy reviews, file delivery, remediation tracking, and general technical support - Track compliance-related requirements for new product and service launches and coordinate action items across stakeholders - Identify, design, and implement AI-driven solutions that improve the efficiency and scalability of compliance and legal operations - Build intelligent systems using large language models, retrieval-augmented generation, agentic workflows, long-term memory, tool use, subagents, and multi-agent architectures for handling legal files - Develop benchmark frameworks and evaluation methodologies, including dataset construction, annotation strategy, and measurement of retrieval quality, latency, groundedness, factuality, and task success - Run independent research and rapidly prototype new ideas for ambiguous legal and compliance workflow challenges - Support related administrative and operational activities in line with internal compliance requirements

Requirements

- Around 5 years of experience in compliance technology, legal technology, regulatory technology, or software development supporting compliance or legal functions - Strong understanding of relevant compliance rules, regulations, and day-to-day compliance operations - Bachelor's degree required, with a background in AI, Computer Science, Engineering, Law, or a related discipline - Production experience with retrieval-augmented generation systems, including embeddings, vector databases, hybrid search, reranking, chunking, text preprocessing, and multimodal parsing - Experience implementing advanced agentic RAG approaches such as Self-RAG, Corrective RAG, adaptive retrieval, and multi-hop reasoning workflows - Strong grasp of LLM and agent fundamentals, including prompt engineering, context engineering, memory, planning, tool use, MCP, and multi-agent systems - Ability to run independent research, solve ambiguous problems, and rapidly prototype practical solutions - Strong written and verbal communication skills

Nice to have

- Bilingual English and Mandarin to coordinate with overseas partners and stakeholders

Benefits and work setup

- Work-from-home arrangement, with flexibility that may vary depending on the business team's needs - Competitive salary and company benefits within a results-driven, global organisation