LLM Data Scientist/Algorithm Engineer (Top Talent Program) (Fully Remote)

Binance; United States; Remote; USD 140,000 - 190,000; Full time

Job summary

Binance is seeking an experienced NLP/LLM Engineer to design and optimize end-to-end LLM pipelines for insightful decision-making. You will build scalable LLM/RAG frameworks for customer service scheduling and collaborate across teams on advanced multi-agent architectures.

The role requires a Master’s in CS/Data Science and 2+ years in NLP/LLM, with strong English and Chinese communication skills to coordinate across diverse teams.

Qualifications

• Master's degree or higher in CS/Data Science with 2+ years of NLP/LLM experience.
• Expertise in prompt engineering, multi-agent workflows, and vLLM/SGLang inference.
• Fluent in English and Chinese for cross-team collaboration.

Responsibilities

• Own the full LLM pipeline from data preparation to production usage.
• Build and maintain RAG QA/search systems and multi-agent LLM architectures.
• Create evaluation pipelines to measure model quality, bias, and hallucination rates.

Skills

NLP/LLM experience

Prompt engineering

Multi-agent workflows

vLLM/SGLang inference

Education

Master's or higher in CS/Data Science

Job description

• Develop and refine LLMs to extract actionable insights and improve decision-making.
• Build scalable LLM/RAG frameworks for customer service scheduling.

Responsibilities:

• Own the full LLM pipeline from data preparation to production usage.
• Build and maintain RAG QA/search systems and multi-agent LLM architectures.
• Create evaluation pipelines to measure model quality, bias, and hallucination rates.

Requirements:

• Master's degree or higher in CS/Data Science with 2+ years of NLP/LLM experience.
• Expertise in prompt engineering, multi-agent workflows, and vLLM/SGLang inference.
• Fluent in English and Chinese for cross-team collaboration.

Binance

Binance is a leading global blockchain ecosystem and the world's largest cryptocurrency exchange by trading volume and registered users. It is trusted by 300+ million users across 100+ countries and maintains a remote-first, flat, results-driven culture.