Senior Machine Learning Data Engineer

Role Direction:

We are looking for a Risk Control Algorithm / ML Expert with 5-7 years of experience to build and optimize machine learning-based risk control models for financial trading scenarios. This role requires combining user behavior data, trading data, fund data, and financial time-series data to identify abnormal transactions, fraud risks, fund risks, account risks, and potential business risks — and to drive model deployment into production systems.

Key Responsibilities:

- Design, develop, train, validate, and deploy risk control algorithm models — covering anomaly detection, risk scoring, fraud identification, user segmentation, trading behavior monitoring, and more

- Build feature systems and model frameworks based on user behavior, trading records, deposit/withdrawal data, market data, financial time-series, and other data sources

- Apply machine learning, deep learning, and statistical modeling methods to improve risk identification accuracy, recall, and model stability

- Build enterprise-grade risk control model pipelines — including data processing, feature engineering, model training, evaluation, deployment, monitoring, and iteration

- Collaborate with Product, Risk Control, Data Engineering, Backend, and Business teams to integrate model outputs into actual risk control strategies and business workflows

- Continuously monitor model performance — analyzing model drift, false positives, false negatives, and anomalous fluctuations — and drive optimization

- Contribute to MLOps process development — improving standardization and automation across model development, deployment, monitoring, and rollback

Required Qualifications:

- 5+ years of experience in data science, machine learning, risk control algorithms, or quantitative modeling — with hands-on experience deploying enterprise-grade ML models

- Proficient in Python — familiar with Pandas, NumPy, Scikit-learn, PyTorch, and other common data analysis and ML toolkits

- Proficient in SQL — able to independently complete complex data extraction, cleaning, feature construction, and validation

- Solid foundation in machine learning and statistical analysis — familiar with classification, regression, clustering, anomaly detection, time-series modeling, and related methods

- Familiar with financial time-series data analysis — able to understand market volatility, trading behavior, fund flows, and other financial data characteristics

- Strong model evaluation skills — familiar with AUC, KS, Precision, Recall, F1, Lift, stability monitoring, and related metrics

- Familiar with MLOps — including model deployment, version management, model monitoring, feature management, automated training, and deployment pipelines

- Strong business understanding — able to break down complex risk control problems into modelable, verifiable, and implementable algorithmic solutions

Preferred Qualifications:

- Experience in FinTech, trading platforms, brokerages, payments, anti-fraud, AML, credit risk, quantitative trading, or risk management

- Experience with real-time risk control, abnormal transaction detection, fund risk monitoring, account risk scoring, or user behavior modeling

- Familiarity with financial trading scenarios — such as deposits, withdrawals, trading behavior, positions, leverage, slippage, rebates, agent/IB systems, and more

- Experience with large-scale data processing — familiarity with Spark, Airflow, Kafka, Feature Store, Docker, Kubernetes, or similar tools is a plus

- Experience building and deploying models from 0 to 1 in production — able to independently drive models from experimentation to production

- Good English proficiency — able to read technical documentation and support cross-team communication

What We're Looking For:

You are not a traditional engineer who just waits for requirements, writes code, and delivers tasks. We want someone who:

- Is keen on new technologies and tools

- Proactively uses AI to improve efficiency

- Solves problems with mature components and platform capabilities

- Cares about product experience and real user value

- Can move fast, launch quickly, and iterate effectively in a small team

- Wants to transform a traditional broker system into a truly trader-centric global trading platform

Traze's goal is not to build another traditional trading entry point, but to create a simpler, more efficient, more capable, and more user-friendly platform for traders around the world.

岗位定位：

我们正在寻找一位具备 5-7 年经验的风控算法模型专家，负责搭建和优化面向金融交易场景的机器学习风控模型。该岗位需要结合用户行为数据、交易数据、资金数据和金融时间序列数据，识别异常交易、欺诈风险、资金风险、账户风险及潜在业务风险，并推动模型在真实业务系统中落地。

岗位职责：

- 负责风控算法模型的设计、开发、训练、验证和上线，包括异常检测、风险评分、欺诈识别、用户分层、交易行为监控等方向

- 基于用户行为、交易记录、入出金数据、市场行情、金融时间序列等数据，构建特征体系和模型框架

- 使用机器学习、深度学习、统计建模等方法，提升风险识别准确率、召回率和模型稳定性

- 搭建企业级风控模型流程，包括数据处理、特征工程、模型训练、模型评估、上线部署、监控和迭代

- 与产品、风控、数据工程、后端开发及业务团队协作，将模型结果接入实际风控策略和业务流程

- 持续监控模型表现，分析模型漂移、误报、漏报和异常波动，并推动模型优化

- 参与 MLOps 流程建设，提升模型开发、部署、监控和回滚的标准化与自动化能力

任职要求：

- 5+年数据科学、机器学习、风控算法或量化建模相关经验，有企业级机器学习模型落地经验

- 熟练使用 Python，熟悉 Pandas、NumPy、Scikit-learn、PyTorch 等常用数据分析和机器学习工具包

- 熟练使用 SQL，能够独立完成复杂数据提取、清洗、特征构建和数据验证

- 具备扎实的机器学习和统计分析能力，熟悉分类、回归、聚类、异常检测、时间序列建模等方法

- 熟悉金融时间序列数据分析，能理解行情波动、交易行为、资金流向等金融数据特征

- 有较强的模型评估能力，熟悉 AUC、KS、Precision、Recall、F1、Lift、稳定性监控等指标

- 熟悉 MLOps，了解模型上线、版本管理、模型监控、特征管理、自动化训练和部署流程

- 具备良好的业务理解能力，能够将复杂风控问题拆解为可建模、可验证、可落地的算法方案

加分项：

- 有金融科技、交易平台、券商、支付、反欺诈、反洗钱、信用风控、量化交易或风险管理经验

- 有实时风控、异常交易识别、资金风险监控、账户风险评分、用户行为建模经验

- 熟悉金融交易场景，如入金、出金、交易行为、持仓、杠杆、滑点、返佣、代理体系等

- 有大规模数据处理经验，熟悉 Spark、Airflow、Kafka、Feature Store、Docker、Kubernetes 等工具优先

- 有模型从 0 到 1 搭建并成功上线的经验，能独立推动模型从实验走向生产

- 英文能力良好，能阅读英文技术文档，并支持跨团队沟通

我们希望你是这样的人：

你不是传统意义上只等需求、写代码、交付任务的研发。我们希望你：

- 对新技术和新工具敏感

- 愿意主动使用 AI 提升效率

- 能用成熟组件和平台能力解决问题

- 关注产品体验和真实用户价值

- 能在小团队里快速推进、快速上线、快速复盘

- 希望把一个传统 broker 系统，升级成真正以交易者为中心的全球化交易平台

Traze 的目标不是再做一个传统交易入口，而是做一个更简单、更高效、更有工具能力、更适合全球交易者使用的平台。