LLM Systems Engineer: Context Pipelines & Quality

LinkPOWER Technology Company Limited; Hong Kong; On-site; HKD 500,000 - 900,000; Full time

Job summary

LinkPOWER Technology Company Limited in Hong Kong seeks an AI/ML Platform Engineer to design and maintain LLM context pipelines, including RAG, memory, and prompt chaining for robust personalisation.

The role focuses on optimizing embeddings, semantic search, and knowledge graphs, extracting and linking data, and developing metadata schemas. Collaboration with ML teams and PMs drives model evolution and quality assurance.

Responsibilities

• Design and maintain LLM context pipelines (RAG, memory, prompt chaining, semantic search).
• Optimize chunking, embeddings, and knowledge structures (taxonomies, ontologies, knowledge graphs).
• Extract and link structured/unstructured data; define metadata schemas and inference rules for personalization.
• Create training data, test plans, cases, and criteria; execute manual and automated tests for APIs, UIs, and integrations.
• Build and maintain test automation integrated into CI/CD.
• Monitor context quality, latency, and hallucinations; perform root cause analysis and drive defect resolution.
• Collaborate with ML engineers, PMs, and dev teams to align requirements and evolve models and test coverage.

Job description

LinkPOWER Technology Company Limited in Hong Kong seeks an AI/ML Platform Engineer to design and maintain LLM context pipelines, including RAG, memory, and prompt chaining for robust personalisation.

The role focuses on optimizing embeddings, semantic search, and knowledge graphs, extracting and linking data, and developing metadata schemas. Collaboration with ML teams and PMs drives model evolution and quality assurance.