
美图反作弊工程师(北京)
任职要求
Direct experience in anti-fraud scenarios for online commu…
工作职责
● 岗位职责 主导搭建并迭代全业务线风控与反作弊体系,设计多场景策略框架,保障数据真实和公司利益 分析爬虫、虚假交易、账号作弊、刷量薅羊毛等风险特征,输出并落地风控策略,持续优化拦截率与误判率 搭建实时风险监测体系,及时发现新型作弊手段,主导重大风险应急处置,降低业务损失 协同研发、运营等跨团队推动风控工具/引擎落地,输出风控知识,提升团队风险防控意识 ● 任职资格 计算机、统计学、数学等相关专业,本科及以上学历,3 年及以上风控反作弊经验 扎实数据分析能力,熟练使用 SQL 进行数据查询与清洗,精通 Python(Pandas、NumPy、Scikit-learn 等库)开展特征工程与模型开发 熟悉风控规则引擎(如 URule、Drools)的使用与配置,能将业务规则转化为可落地的风控策略 熟悉 CDN/WAF、设备指纹、浏览器指纹等反爬技术及对抗策略,了解常见爬虫攻击原理与防御逻辑 熟练使用逻辑回归、决策树、随机森林、XGBoost/LightGBM 等风控常用模型,理解模型训练、评估与线上推理全流程 具备敏锐的风险洞察力与逻辑分析能力,优秀的跨团队协作与项目推动能力,能够在高压环境下快速响应并解决复杂风险问题 ● 加分项(选填) 有社区、订阅场景反作弊经验优先 Job Summary We are seeking a highly data mining and technically proficient Anti-Fraud Engineer to join our Risk Control team. In this role, you will be at the forefront of our defense system, designing scalable frameworks to combat sophisticated threats such as bot attacks, fraudulent transactions, and incentive abuse. You will leverage data science, rule engines, and adversarial technologies to safeguard our ecosy Job Responsibilities System Architecture: Lead the end-to-end design and iteration of the company-wide risk control and anti-fraud framework. Architect multi-scenario strategies to safeguard data integrity and corporate interests. Risk Mitigation: Analyze behavioral signatures of risks such as web crawlers, fraudulent transactions, account abuse, and incentive scalping. Develop and deploy targeted strategies to continuously optimize detection precision and minimize false positive/negative rates. Real-time Monitoring & Response: Build real-time risk monitoring systems to identify emerging fraud patterns. Spearhead emergency response efforts for major risk incidents to minimize business losses. Cross-functional Collaboration: Partner with R&D, Operations, and other cross-functional teams to implement risk engines and tools. Disseminate expertise to enhance overall risk awareness across the organization. Job Requirements Educational Background: Bachelor’s degree or above in Computer Science, Statistics, Mathematics, or a related field, with 3+ years of professional experience in risk control and anti-fraud. Data Proficiency: Strong data analysis skills with proficiency in SQL for data extraction and cleaning. Expert in Python (Pandas, NumPy, Scikit-learn) for advanced feature engineering and model development. Rule Engine Expertise: Familiar with the configuration and deployment of risk rule engines (e.g., URule, Drools), with the ability to translate complex business logic into executable technical strategies. Adversarial Defense: Well-versed in anti-crawling technologies and adversarial countermeasures, including CDN/WAF, device fingerprinting, and browser fingerprinting; deep understanding of crawler attack vectors and defense mechanisms. Machine Learning: Hands-on experience with mainstream risk models (e.g., Logistic Regression, Decision Trees, Random Forest, XGBoost/LightGBM) and a thorough understanding of the full lifecycle: training, evaluation, and online inference. Soft Skills: Acute risk insight and logical reasoning. Excellent cross-team collaboration and project management skills, with the ability to resolve complex issues under high-pressure environments. Preferred
1. 构建基于机器学习或深度学习模型的在线识别系统,保证极高精准和召回的前提下,实时处置无效流量或风险物料。 2. 建设强大的离线挖掘平台,使用多种无监督、半监督手段帮助算法和运营同学探索新的异常pattern。 3. 研发风控产品平台,服务阿里妈妈和淘天集团更多相关的业务,不断提升迭代效率。每年为广告主挽回损失过百亿元,同时为平台拦截百亿以上的风险物料。
1. 构建基于机器学习或深度学习模型的在线识别系统,保证极高精准和召回的前提下,实时处置无效流量或风险物料。 2. 建设强大的离线挖掘平台,使用多种无监督、半监督手段帮助算法和运营同学探索新的异常pattern。 3. 研发风控产品平台,服务阿里妈妈和淘天集团更多相关的业务,不断提升迭代效率。每年为广告主挽回损失过百亿元,同时为平台拦截百亿以上的风险物料。
【关于我们】 滴滴自2018年初收购巴西出行平台99开始,正式启动了国际化战略。目前,滴滴的国际业务覆盖了拉美、亚太和非洲的14个国家,为当地市场提供以出行为主、涵盖外卖和金融的多样化服务。 岗位职责: 1. 深入理解滴滴海外出行的业务模式、流程和系统架构,发现潜在作弊风险点,和产品技术、业务运营高效沟通,设计合理的风控策略架构; 2. 独立思考业务场景可能存在的作弊风险,设计合理的数据埋点体系,通过大数据分析,定量识别潜在的风险和业务影响; 3. 与相关团队紧密配合,通过大数据挖掘,找到作弊者的行为特点,快速形成有效的打击策略,持续迭代优化某个业务或场景的风控效果; 4. 针对某个业务或场景建立合理的指标体系,在对抗过程中不断完善监控体系,与数据工程团队配合,形成可视化的监控系统快速发现作弊。
【关于我们】 滴滴自2018年初收购巴西出行平台99开始,正式启动了国际化战略。目前,滴滴的国际业务覆盖了拉美、亚太和非洲的14个国家,为当地市场提供以出行为主、涵盖外卖和金融的多样化服务。 岗位职责: 1. 深入理解滴滴海外出行的业务模式、流程和系统架构,发现潜在作弊风险点,和产品技术、业务运营高效沟通,设计合理的风控策略架构; 2. 独立负责账号安全,设计合理的数据埋点体系,通过大数据分析,定量识别潜在的盗号风险和业务影响; 3. 与相关团队紧密配合,通过大数据挖掘,找到作弊者的行为特点,快速形成有效的打击策略,持续迭代优化账号安全的风控策略; 4. 建立合理的指标体系,在对抗过程中不断完善监控体系,与数据工程团队配合,形成可视化的监控系统快速发现作弊。