LLM Data Scientist / Algorithm Engineer (Top Talent Program)

About Binance

Binance stands as a premier global blockchain platform, hosting the world's largest cryptocurrency exchange by trading volume and registered users. Trusted by over 300 million users across more than 100 countries, Binance excels in security, transparent user fund management, rapid trading engines, deep liquidity, and a comprehensive array of digital-asset offerings ranging from trading, finance, education, research, payments to institutional services. The company is dedicated to broadening financial freedom and access worldwide.

Role Overview

We are looking for a highly competent individual to join our cutting-edge AI team, concentrating on pioneering large language model (LLM) technologies and solutions. The chosen candidate will concentrate on developing, enhancing, and deploying LLMs to decode actionable insights that drive business decisions and elevate prompt accuracy. This role entails delivering scalable, resilient frameworks like LLM and Retrieval-Augmented Generation (RAG) customized for customer service scheduling, fostering innovation, and preserving a strong competitive stance in the market.

Key Responsibilities

• Manage the entire lifecycle of LLM implementation, from data preparation through to real-world production deployment.
• Devise, refine, and enhance various prompt formats including zero-/few-shot, chain-of-thought, tool-calling to improve model effectiveness and ensure safety across different products and languages.
• Develop and sustain Retrieval-Augmented Generation QA/search systems integrating multi-source knowledge repositories.
• Demonstrate proficiency with vLLM and SGLang inference structures, and have experience deploying and maintaining LLM services on multi-GPU or clustered setups.
• Design, implement, and operate multi-agent LLM architectures such as LangGraph, CrewAI, and AutoGen involving task breakdown, agent coordination, memory management, and tool-calling workflows.
• Create evaluation systems employing automatic metrics and human feedback to assess prompt/model performance, bias, and hallucination tendencies.
• Work collaboratively with product and customer service units to embed AI models within conversational chatbot use cases.
• Stay updated on the latest research trends, publish technical blogs, and continuously refine existing architectures.

Qualifications & Experience

• Possess a Master’s degree or higher in Computer Science, Data Science, or related disciplines.
• Bring at least 2 years of experience in deep learning and natural language processing, with over 1 year specifically working with LLMs involving supervised fine-tuning, direct policy optimization, retrieval-augmented generation, quantization, and inference enhancements.
• Exhibit expertise in prompt engineering and tuning techniques, including few-shot design, structured prompts, prefix/p-tuning, reward re-ranking, and safety filtering strategies.
• Have practical knowledge in constructing and deploying multi-agent LLM workflows with understanding of orchestrator roles, shared memory usage, prolonged planning, and guardrail mechanisms.
• Fluency in English and Chinese to ensure effective communication across teams.

Employee Benefits

• Opportunity to shape the future landscape of finance.
• Collaborate with top-tier talent in a globally integrated, flat organizational structure.
• Engage with challenging projects in a fast-moving, innovative environment with significant autonomy.
• Work in a result-oriented culture that promotes career advancement and ongoing skill development.
• Access competitive compensation packages along with comprehensive company benefits.
• Enjoy a remote-first work setup, allowing you to work entirely from home, subject to role and team specifics.

Diversity & Equal Opportunity

Binance champions equal opportunity employment and recognizes diversity as an essential component of our success.

Use of AI in Recruitment

In our hiring process, AI tools may assist by screening applications, analyzing resumes, and assessing responses to identify irregularities or validate materials. These technologies support but do not replace human judgment. Final employment decisions remain the responsibility of our recruitment team. Candidates seeking further details about data processing may contact us.