Binance is a leading global blockchain ecosystem behind the world’s largest cryptocurrency exchange by trading volume and registered users. We are trusted by 300+ million people in 100+ countries for our industry-leading security, user fund transparency, trading engine speed, deep liquidity, and an unmatched portfolio of digital-asset products. Binance offerings range from trading and finance to education, research, payments, institutional services, Web3 features, and more. We leverage the power of digital assets and blockchain to build an inclusive financial ecosystem to advance the freedom of money and improve financial access for people around the world.
You will focus on user-facing financial AI and equity research scenarios, identifying and developing data and knowledge algorithms that are strategically valuable for Binance to build over the long term. By combining large language models, natural language processing, machine learning, and reinforcement learning, you will enable the system to better understand user queries, recognize financial entities and temporal information, retrieve timely and relevant information, and generate results that are measurable and continuously optimizable. You will own the full lifecycle, from problem definition and data development to model training and production evaluation.
Responsibilities
Identify high-value financial data algorithm problems that are worth building in-house, and evaluate the effectiveness, cost, and long-term maintainability of different approaches, including external data sources, rule-based processing, traditional models, and large language model solutions.
Design, train, evaluate, and optimize financial data and knowledge algorithms in production, covering areas such as user query understanding, document understanding, information extraction, entity recognition and linking, event detection, timeliness assessment, classification and tagging, deduplication and consolidation, and quality scoring.
Develop multi-channel retrieval, relevance modeling, and financial ranking algorithms that dynamically balance relevance, timeliness, source authority, popularity, content quality, and other domain-specific financial signals based on user queries.
Build training datasets, labeling systems, and evaluation benchmarks for market data, fundamentals, earnings reports, announcements, news, research reports, and licensed investment research data, while addressing sample bias, label noise, source conflicts, and market changes.
Select and optimize the appropriate methods for each task, including large language models, NLP models, multimodal models, graph algorithms, traditional machine learning, or rule-based approaches, balancing accuracy, recall, explainability, timeliness, and cost. Collaborate with Financial AI Engineers to integrate algorithms into a unified knowledge processing and retrieval pipeline and deploy them reliably into production.
Apply supervised fine-tuning, reinforcement learning, preference optimization, active learning, or semi-supervised learning, leveraging expert feedback and production data to continuously improve data processing models and financial data agents.
Establish both offline and online evaluation frameworks to measure accuracy, recall, ranking quality, timeliness, irrelevant information ratio, coverage, consistency, and cross-market generalization. Evaluate the authority relationship between tool usage and retrieval-augmented generation, ensure that historical evidence does not override updated facts, and attribute errors across data, retrieval, ranking, and model layers.
Requirements
Experience in financial data, brokerage, trading platforms, research institutions, wealth management, or fintech-related algorithm development.
Experience in extracting, linking, event detection, or quality evaluation for financial content such as earnings reports, announcements, research reports, and news.
Experience in financial large model post-training, reinforcement learning, knowledge graphs, multimodal document understanding, or data agent optimization.
Experience with active learning, weak supervision, human feedback loops, or large-scale data labeling and evaluation systems.
Experience in cross-market or cross-language model transfer, or in conducting independent evaluation and calibration for different markets.
