Senior LLM Engineer: RAG & Agentic Workflows Architect

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

Job summary

Randstad Hong Kong Limited seeks a hands-on ML engineer to design and develop LLM-powered applications, including multi-agent systems and agentic workflows with prompt orchestration and tool calling. You will build robust RAG capabilities, covering data ingestion, embedding, vector search, reranking, and citation attribution.

You will collaborate with data scientists and software engineers to deploy scalable AI solutions in a cloud-native environment, focusing on performance, reliability, and

Qualifications

• Minimum 3 years of hands-on experience building and deploying production-grade software services and LLM applications.
• Strong Python proficiency with production-grade APIs, asynchronous jobs, and CI/CD pipelines.
• Experience with modern LLM frameworks and orchestration tools (e.g., LangChain, LangGraph, LlamaIndex).
• Experience with RAG tuning, prompt engineering, and PEFT/LoRA.
• Knowledge of evaluating groundedness, faithfulness, and running A/B tests.
• Experience deploying AI models in production using containers, cloud infra, and scalable architectures.

Responsibilities

• Design and develop LLM-powered applications, including multi-agent systems and agentic workflows with prompt orchestration and tool calling.
• Build RAG capabilities: data ingestion, chunking, embedding, vector search, reranking, citation attribution.
• Develop and maintain data pipelines governing external knowledge sources and ensuring freshness and access control.
• Improve model performance via prompt optimisation, dataset curation, and lightweight adaptation (LoRA/PEFT).
• Implement LLMOps practices: deployment, observability, evaluation, incident response, latency/cost optimisation.
• Collaborate with ML engineers and software developers to integrate LLM systems via APIs, microservices, and platforms.
• Track advances in LLM ecosystem and translate techniques into measurable production improvements.

Skills

Python

LLM apps

RAG

LangChain

LangGraph

LlamaIndex

Vector search

Prompt optimization

Tool calling

Multi-agent systems

Agentic workflows

LoRA/PEFT

CI/CD

Kubernetes

Cloud deployments

Education

Bachelor Degree

Tools

OpenTelemetry

Job description

Randstad Hong Kong Limited seeks a hands-on ML engineer to design and develop LLM-powered applications, including multi-agent systems and agentic workflows with prompt orchestration and tool calling. You will build robust RAG capabilities, covering data ingestion, embedding, vector search, reranking, and citation attribution.

You will collaborate with data scientists and software engineers to deploy scalable AI solutions in a cloud-native environment, focusing on performance, reliability, and