Frontier AI Agent Engineer

Binance; Hong Kong; On-site; HKD 900,000 - 1,300,000; Full time

Benefits offered by this job

Work-from-home arrangement

Competitive salary

Career growth

Continuous learning

Job summary

Binance is seeking a research-minded engineer to join the AI Infra team. You will work at the intersection of frontier model capabilities and real-world agent deployment, designing adaptive retrieval pipelines and multi-agent systems while collaborating with researchers and engineers.

This role blends research and engineering, requiring you to generate ideas, run experiments, ship prototypes, and iterate rapidly based on user feedback.

Qualifications

• Experience with production-grade LLM and AI agent systems.
• Hands-on with retrieval pipelines and vector stores (Qdrant, Milvus, Pinecone, Weaviate).
• Strong understanding of tokenization, chunking, and data parsing for RAG.
• Experience with Agent Harness runtimes and sandboxing.
• Familiarity with LLM APIs, memory, and tool-calling patterns.
• Proven ability to translate ideas into runnable prototypes with fast iteration.

Responsibilities

• Design and operate next-gen retrieval pipelines and adaptive multi-hop retrieval workflows.
• Collaborate with researchers to implement model-capability innovations including context management and multi-agent architectures.
• Define evaluation methodologies and construct benchmarks for agent intelligence and retrieval latency.
• Design experiments using real-world feedback to continuously improve agent and retrieval performance.

Skills

LLM & Agent Fundamentals

Agent Harness Engineering

RAG & Agentic RAG

Context Engineering

Memory & Long-term

Multi-Agent Systems

Prompt Engineering

Tools

Qdrant

Milvus

Pinecone

Weaviate

Pi Agent

AgentScope 2.0

Kubernetes

Job description

Binance is seeking a research-minded engineer to join the AI Infra team. You will work at the intersection of frontier model capabilities and real-world agent deployment, designing adaptive retrieval pipelines and multi-agent systems while collaborating with researchers and engineers.

This role blends research and engineering, requiring you to generate ideas, run experiments, ship prototypes, and iterate rapidly based on user feedback.