AI Engineer Intern (RAG/LLM)

ICC (Hong Kong) Limited; Hong Kong; On-site; HKD 900,000 - 1,300,000; Full time

Job summary

ICC (Hong Kong) Limited is seeking a versatile engineer to lead full-stack development, building features with React and FastAPI, and integrating LLM components.

You will design and optimize RAG pipelines for contract processing, including document parsing (PDFs, OCR, tables), embeddings, and retrieval, with emphasis on reliability.

Qualifications

• Strong Python proficiency and API-focused backend skills.
• Full-stack capability with React/TypeScript and FastAPI.
• Genuine understanding or experience with LLMs or RAG systems.
• Ability to turn ideas into working prototypes independently.

Responsibilities

• Contribute to end-to-end development (frontend + backend) with React + FastAPI.
• Design and optimize RAG pipelines for contract use cases (document parsing, embedding, retrieval).
• Engineering challenges: OCR, layout analysis, table extraction, robust LLM systems.
• Develop LLM application features including prompts, context management, and traceability.

Skills

Python

Full-stack (React/TS)

LLMs / RAG

Prototype development

Strong problem solving

Tools

LangChain

LlamaIndex

Milvus

pgvector

Pinecone

vLLM

Ollama

Job description

• Full-Stack System DevelopmentContribute to end-to-end development (frontend + backend)
  Build features using React + FastAPI
• RAG System Design & OptimizationBuild and improve RAG pipelines for contract use cases:
  Document parsing (PDFs, OCR, tables)
  Chunking, embedding, and retrieval optimization
  Long-context reasoning and context management
  Improve retrieval accuracy and generation reliability
• LLM Application EngineeringDesign prompt and context management strategies
  Reduce hallucinations in high-stakes outputs
  Build traceable and auditable AI outputs
• Engineering ChallengesComplex document processing (OCR, layout analysis, table extraction)
  Structured extraction from unstructured contract data
  Improve robustness of LLM systems in real-world scenarios

Job requirements:

Must have:

• Strong proficiency in Python
• Basic full-stack capability (React / TypeScript / FastAPI)
• Genuine understanding or experience with LLMs or RAG systems
• Ability to turn ideas into working prototypes independently
• Strong logical thinking and engineering mindset

Nice to have:

• Experience with LangChain / LlamaIndex
• Familiarity with vector databases (Milvus / pgvector / Pinecone)
• Experience with local model deployment (vLLM / Ollama)
• Experience with OCR / document understanding pipelines
• Interest in long-context reasoning or retrieval optimization
Experience using AI coding tools (Claude Code, Codex, etc.)