通信,拉近世界的距离!
您的位置:首页 >资讯 >业界新闻 >查看资讯

让 AI 可信:Anthropic 工程师 Saran 的数据治理之路

移动通信网 程然

当 Claude 在几秒内回答一个复杂问题时,它给出的信息从何而来?答案是否可靠、是否存在隐私风险?这是 Anthropic 工程师 Gurmeet Saran 每天都在面对的问题。作为这家位于旧金山的 AI 研究公司的技术团队成员,同时也是 IEEE 高级会员,Saran 正在构建一套基础设施,用来理解聊天机器人行为背后的机制,并确保其底层架构能够产出可靠输出。“我的工作是确保数据准确、结构化且可信,同时保证私有数据始终私密,”他表示,“隐私不能事后叠加,必须从一开始就是架构的一部分。”

安全优先:LLM 可观测性仍是难题

安全性是 Anthropic 的核心优先级。根据 Claude 的使命声明,公司希望这一 AI 智能体“对与其协作或为其服务的人以及社会真正有用,同时避免不安全、不道德或欺骗性的行为”。实现这一使命的关键,在于让研究者能够观察和理解 AI 模型的行为方式——这对大语言模型而言尤其棘手。Saran 直言:“我们有成熟的工具来监控数据库和 Web 服务,但对 LLM 来说,一切仍在探索阶段。”这一表态揭示了当前 AI 行业面临的真实处境:模型的复杂度已经超出了现有可观测性工具的能力边界,而信任与安全恰恰建立在这一基础之上。如果无法信任底层逻辑,就无法对模型行为做出可靠判断,这会带来现实的安全影响。

数据治理的三段历练

Saran 的职业轨迹横跨多家科技公司,每一段经历都在回答同一个问题:如何让数据可信。他的技术兴趣始于加尔各答的童年时代,当时他尽可能泡在学校机房里。“第一次使用计算机时,让我震撼的是你写下指令,机器会精确执行,没有任何歧义,那对我来说像是魔法。”2004 年,他在 Kalyani 大学获得电子与通信工程学士学位,2010 年在 Birla Institute of Technology and Science 取得软件工程硕士学位。

2007 年,还是研究生的他加入雅虎班加罗尔团队担任数据组技术负责人。彼时他遇到了塑造其职业方向的问题:不同团队用略有差异的逻辑计算同一个网站指标,比如日活跃用户数,最终得到相互矛盾的结果。“听起来是个小问题,但当管理层基于这些数据做资源分配和产品决策时,不知道哪个数字正确就成了大麻烦。”团队为此构建了一个带有统一定义指标的中央数据仓库,充当单一事实来源。Saran 设计的模型精确定义了“日活跃用户”的含义,明确了哪些事件被计入、边界情况如何处理。在此基础上,团队打造了内部网络分析工具 Yahoo! Digits,服务于公司全线产品。

2010 年,Saran 转投 Intuit 担任资深工程师,参与构建统一 TurboTax 和 QuickBooks 等产品数据的分析基础设施。任职期间,Intuit 开发了实时检测欺诈性纳税申报的功能,工程团队需要在严格遵守美国国税局政策的前提下构建保护纳税人数据的系统。“这让我明白,隐私和治理不能事后附加到软件上,”他说,“它们必须从一开始就融入架构。”这段经历也成为他如今在 Anthropic 工作的方法论基石。

架构问题而非编码缺陷

2016 年,Saran 加入 Meta(当时的 Facebook)担任首席软件工程师。彼时公司正深陷视频指标争议——Meta 公开承认一个错误导致平台夸大了视频的观看次数。“这不是编码 Bug,而是架构问题,”Saran 评述道。他在 Meta 构建的系统延续了在 Intuit 的思路:通过单一真实数据源消除指标混乱。这一判断在今天看来颇具预见性——当 AI 模型的输出被广泛用于决策时,数据管道中任何环节的架构缺陷都可能被放大为系统性风险。

全球工程师社区的价值

作为 IEEE 高级会员,Saran 也看重专业组织在 AI 时代的连接作用。“IEEE 在许多国家开展活动,那里的工程师缺少进入全球专业社区的便捷通道。随着 AI 开发扩展到少数机构之外,这种连接变得重要。”在 AI 人才和研发高度集中于头部企业的当下,跨地域、跨组织的专业协作机制,或许正是行业建立统一可信标准的必要条件。

从雅虎的指标统一,到 Intuit 的隐私架构,再到 Meta 的数据治理,Saran 近二十年的职业生涯始终围绕一个主题:让数据值得信任。如今他把这一命题带到了生成式 AI 的最前沿。在 LLM 可观测性工具尚不成熟的现阶段,这种“从架构层面内建可信度”的工程思路,为行业提供了一个务实的技术参照。


出处:Meet the Anthropic Engineer Keeping Its AI Agent Trustworthy

英文原文
If you’ve ever wondered how Claude and similar AI agents can answer complex questions in seconds and where they get the information, you’re not alone. Gurmeet Saran works to find that answer every day as an engineer at Anthropic in San Francisco, the AI research company behind Claude. The IEEE senior member is developing an infrastructure to understand the mechanics behind the chatbot’s behavior and to ensure its underlying architecture produces reliable outputs. Gurmeet Saran Employer Anthropic in San Francisco Title Member of technical staff Member grade IEEE senior member Alma maters University of Kalyani, in India; Birla Institute of Technology and Science, in Pilani, India “My job is to make sure the data is accurate, structured, and trustworthy, and that private data stays private,” he says. “As I learned earlier in my career, privacy can’t be a layer you bolt on—it has to be part of the architecture from the start.” If you can’t trust the logic, he adds, you can’t make reliable claims about the behavior, and that has real safety implications. Safety is a priority for Anthropic, he says. According to Claude’s mission statement , the company wants the AI agent to be “genuinely helpful to the people it works with or on behalf of, as well as to society, while avoiding actions that are unsafe, unethical, or deceptive.” Crucial to the mission is ensuring that researchers can observe and understand how AI models behave, a challenge that is especially acute for large language models , Saran says. “We have decent tools to monitor databases and web services,” he says, “but for LLMs, we’re still figuring it out.” Solving data quality problems at Yahoo, Intuit, and Meta Saran’s interest in computing systems dates back to his childhood in Kolkata, where he logged as much time as he could in the school computer lab. “I remember the first time I used one of the computers,” he says. “What struck me was that you could write instructions, and the machine would follow them exactly. No ambiguity. “That felt magical to me.” After graduating from high school in 1999, he stayed in India to pursue advanced degrees. He earned a bachelor’s degree in electronics and telecommunication engineering in 2004 at the University of Kalyani . In 2010, he earned a master’s degree in software engineering from the Birla Institute of Technology and Science in Pilani. While a graduate student, he began his career in 2007 at Yahoo as a technical lead of the data team in Bengaluru, India. While there, he encountered the project that shaped his career trajectory: working on web analytics and business intelligence reporting solutions. “We kept running into a frustrating problem,” he recalls. “Different teams would calculate the same website metric—something like daily active site users—using slightly different logic, and end up with conflicting answers. It sounds like a small issue, but when leadership is making decisions about resource allocation and product development using that data, not knowing which number is correct becomes a problem.” The team built a central data repository with defined metrics, essentially serving as a single source of truth. Saran created a model that defined what “daily active site users” means, breaking down which events counted, how to handle edge cases (or when standard code fails), and more. With the new model, data teams across the company could rely on a standard method to determine the number of daily active users for the projects and sites they were responsible for and obtain comparable results using the same logic. From there, Saran and his team built Yahoo! Digits , an internal web analytics tool used for all the company’s products. Saran left Yahoo in 2010 to join Intuit as a staff engineer in Mountain View, Calif. He helped build the analytics infrastructure that unified data across the company’s products, including TurboTax and QuickBooks accounting software. During his tenure, Intuit developed a feature that detected fraudulent tax filings in real time. It required Saran and other engineers to build a system to protect taxpayer data while adhering to strict U.S. Internal Revenue Service policies. “IEEE operates in a lot of countries where engineers don’t have easy pathways into the global professional community. That kind of connection matters as AI development spreads beyond a handful of organizations.” “This taught me that privacy and governance can’t be layers you bolt onto software after the fact,” he says. “They have to be part of the architecture from the start.” Saran joined Meta (known as Facebook at the time) in 2016 as a lead software engineer in Menlo Park, Calif. The company was facing controversy over its video metrics: Meta publicly acknowledged an error that caused it to overstate the number of views for posted videos . “This wasn’t a coding bug; it was an architectural problem,” Saran says. He built a system that worked similarly to what he had developed at Intuit, whereby one single tr
微信扫描分享本文到朋友圈
扫码关注5G通信官方公众号,免费领取以下5G精品资料
5G通信公众号
  • 1、回复“YD5GAI”免费领取《中国移动:5G网络AI应用典型场景技术解决方案白皮书》
  • 2、回复“5G6G”免费领取《5G_6G毫米波测试技术白皮书-2022_03-21》
  • 3、回复“YD6G”免费领取《中国移动:6G至简无线接入网白皮书》
  • 4、回复“LTBPS”免费领取《《中国联通5G终端白皮书》》
  • 5、回复“ZGDX”免费领取《中国电信5GNTN技术白皮书》
  • 6、回复“TXSB”免费领取《通信设备安装工程施工工艺图解》
  • 7、回复“YDSL”免费领取《中国移动算力并网白皮书》
  • 8、回复“5GX3”免费领取《R1623501-g605G的系统架构1》
  • 本周热点本月热点
    通信人才网

    最热通信招聘

    更多