Large Language Model Algorithm Engineer

Calling top-tier AI and Web3 innovators from around the globe! Binance is building the world’s foremost crypto and Web3 platform—and we’re just getting started. We’re assembling a lean, elite LLM Algorithm & Data Science Team (3–5 members) to pioneer next-generation AI solutions in finance, blockchain, and beyond. If you’re young, passionate, and driven by cutting-edge research, this is your moment to make history.

What You’ll Build:

We’re not here to retrofit yesterday’s models—we’re here to invent tomorrow’s:

• Web3 Data Lake LLM: Ingest and reason over massive on-chain and off-chain datasets to power real-time insights.
• Trading System LLM: Design intelligent trading agents, signal generators, and strategy-optimization engines.
• Compliance and Regulatory LLM: Automate compliance, travel-rule enforcement, and on-chain forensics at scale.
• …and beyond: From customer-service agents and personalized navigation experiences to completely new AI-powered Web3 applications.

Your creations will underpin multiple internal products—customer support bots, trading-signal dashboards, compliance workflows, user-experience agents—while also bootstrapping entirely new AI-Web3 innovations.

Responsibilities:

• Model R&D: Architect and train custom foundation LLMs optimized for trading, compliance, and Web3 data.
• Agent Development: Build next-gen multi-agent systems that blend search, RL, and LLM reasoning for real-world tasks.
• Data Engineering: Collaborate on Web3 data-lake design—ETL pipelines, vector stores, real-time feeds.
• Cross-Functional Impact: Partner with product, compliance, and customer-support teams to deploy models in production.
• Innovation Pipeline: Surface new AI-Web3 use cases, prototype MVPs, and shepherd them toward scale.

Requirement:

• Academic Excellence: Master’s or PhD in ML, AI, Computer Science, Mathematics, or related field.
• Research Prodigy: Published in top conferences/journals (ICLR, NeurIPS, ACL, etc.) on foundational LLMs, agent frameworks, or reasoning architectures.
• Reinforcement & Reasoning: Deep expertise in reinforcement learning, reasoning algorithms, or related paradigms—bonus if you’ve built or extended agentic frameworks.
• Polyglot Coder: Fluent in Python; experience with PyTorch, TensorFlow, JAX, or similar. You can prototype new model architectures end-to-end.
• Crypto & Web3 Curious: Enthusiasm for finance engineering, blockchain protocols, on-chain data, DeFi mechanics, or regulatory landscapes.