Agent & Prompt Engineer - AI, Strategy & Operations

The AI Strategy & Operations team sits within Shopee Malaysia Operations and drives AI transformation across the business. We build and operate AI agents, tooling, and automations that take on real operational work, from customer service and return/refund chatbots to fraud, reporting, and process-automation workflows. We partner closely with Regional and Local teams to ship scalable impact fast.

Our mandate is operational excellence. We start where volume and customer impact are highest, in Customer Experience (Customer Service and Return/Refund), then extend what works to the wider Operations org and across the broader organization. The team both builds (production agents and automations) and drives (strategy, change management, and adoption).

Job Description

• As an Agent & Prompt Engineer, you will design, build, and own LLM-powered agents that run in production. Your responsibilities will include:
• Build, maintain, and improve production AI chatbots and agents (e.g. the customer-service buyer bot), owning prompt design, tool/function-calling, and conversation flows
• Translate business problems into agentic workflows — decompose tasks, wire up tools, APIs, and knowledge bases, and orchestrate multi-step agents
• Run rapid prompt experiments and build evals; measure quality, iterate, and ship improvements
• Integrate LLM agents into existing operational workflows (support, fraud review, reporting) to automate manual work
• Partner with PM, data, and ops stakeholders to scope use cases, prototype POCs, and move winners into production
• Monitor live agents, debug failures, and continuously raise reliability and guardrails

• Bachelor's degree or higher in Computer Science, Engineering, Data Science, or a related field — or equivalent demonstrated ability
• Hands-on experience building with LLM APIs: prompt engineering, tool/function-calling, RAG, or agent frameworks
• Proficient in Python and comfortable working with APIs to build and ship integrations end-to-end
• Proven ability to ship — production work, hackathon wins, or personal builds (please share GitHub, demos, or write-ups)
• AI-native mindset: fluent with modern AI tooling and fast at turning ideas into working prototypes
• Bias for action, comfort with ambiguity, and clear communication
• Nice to have: experience with chatbots / CS automation, eval frameworks, or fraud/risk workflows