This Anthropic Engineer Keeps Its AI Agent Trustworthy

“My job is to ensure the info is correct, structured, and reliable, and that personal knowledge stays personal,” he says. “As I realized earlier in my profession, privateness can’t be a layer you bolt on—it must be a part of the structure from the beginning.” If you happen to can’t belief the logic, he provides, you possibly can’t make dependable claims in regards to the conduct, and that has actual security implications.

Security is a precedence for Anthropic, he says. In response to Claude’s mission statement, the corporate needs the AI agent to be “genuinely useful to the folks it really works with or on behalf of, in addition to to society, whereas avoiding actions which are unsafe, unethical, or misleading.” Essential to the mission is making certain that researchers can observe and perceive how AI models behave, a problem that’s particularly acute for large language models, Saran says.

“We have now respectable instruments to watch databases and net companies,” he says, “however for LLMs, we’re nonetheless figuring it out.”

Fixing knowledge high quality issues at Yahoo, Intuit, and Meta

Saran’s curiosity in computing methods dates again to his childhood in Kolkata, the place he logged as a lot time as he may within the faculty laptop lab.

“I keep in mind the primary time I used one of many computer systems,” he says. “What struck me was that you possibly can write directions, and the machine would comply with them precisely. No ambiguity.

“That felt magical to me.”

After graduating from highschool in 1999, he stayed in India to pursue superior levels. He earned a bachelor’s diploma in electronics and telecommunication engineering in 2004 on the University of Kalyani. In 2010, he earned a grasp’s diploma in software engineering from the Birla Institute of Technology and Science in Pilani.

Whereas a graduate scholar, he started his profession in 2007 at Yahoo as a technical lead of the info crew in Bengaluru, India. Whereas there, he encountered the challenge that formed his profession trajectory: engaged on net analytics and enterprise intelligence reporting options.

“We saved working right into a irritating downside,” he remembers. “Totally different groups would calculate the identical web site metric—one thing like every day energetic website customers—utilizing barely completely different logic, and find yourself with conflicting solutions. It seems like a small problem, however when management is making choices about useful resource allocation and product improvement utilizing that knowledge, not understanding which quantity is right turns into an issue.”

The crew constructed a central knowledge repository with outlined metrics, primarily serving as a single supply of fact. Saran created a mannequin that outlined what “every day energetic website customers” means, breaking down which occasions counted, tips on how to deal with edge instances (or when normal code fails), and extra.

With the brand new mannequin, knowledge groups throughout the corporate may depend on a normal technique to find out the variety of every day energetic customers for the tasks and websites they had been chargeable for and acquire comparable outcomes utilizing the identical logic.

From there, Saran and his crew constructed Yahoo! Digits, an inner net analytics instrument used for all the corporate’s merchandise.

Saran left Yahoo in 2010 to affix Intuit as a workers engineer in Mountain View, Calif. He helped construct the analytics infrastructure that unified knowledge throughout the corporate’s merchandise, together with TurboTax and QuickBooks accounting software program. Throughout his tenure, Intuit developed a function that detected fraudulent tax filings in actual time. It required Saran and different engineers to construct a system to guard taxpayer knowledge whereas adhering to strict U.S. Internal Revenue Service insurance policies.

“IEEE operates in loads of nations the place engineers don’t have straightforward pathways into the worldwide skilled neighborhood. That form of connection issues as AI improvement spreads past a handful of organizations.”

“This taught me that privateness and governance can’t be layers you bolt onto software program after the very fact,” he says. “They must be a part of the structure from the beginning.”

Saran joined Meta (referred to as Facebook on the time) in 2016 as a lead software program engineer in Menlo Park, Calif. The corporate was going through controversy over its video metrics: Meta publicly acknowledged an error that brought about it to overstate the number of views for posted videos.

“This wasn’t a coding bug; it was an architectural downside,” Saran says. He constructed a system that labored equally to what he had developed at Intuit, whereby one single fact was outlined to manipulate all video view metrics throughout the location.

That system, Metric360, ultimately turned a part of the location’s structure. By the point he left Meta in 2021, he says, Metric360 was getting used throughout the corporate to calculate video metrics.

He left the corporate to turn into a principal software program engineer at Wealthfront, a monetary companies firm in Palo Alto, Calif. He helped construct the corporate’s knowledge infrastructure.

Constructing reliable AI methods at Anthropic

Saran joined Anthropic in 2025 as a member of the technical workers.

To explain the implications of his work, he makes use of a medical analogy: “Think about a health care provider is monitoring a affected person’s vital signs and making remedy choices primarily based on these numbers. What if one or a number of of the medical devices used had been barely miscalibrated? Even when most readings are high-quality, it’s a significant issue as a result of the tough choices occur within the margins.

“It’s the identical with AI methods. My job is basically to ensure the devices are giving correct readings—that the info means what it says it means.”

His job comes with excessive stakes and complicated issues to unravel.

“AI agents are particularly difficult to look at,” he says. “When a system is taking sequences of actions [generating conversational responses with different users], you want a method to seize what it did, what info it was working from, and whether or not outcomes matched what the tip person anticipated. And it’s a must to do all of that with out exposing anybody’s personal info—the methods are designed to acknowledge delicate knowledge and maintain it protected because it flows by way of. It’s a way more demanding observability downside than something I’ve labored on earlier than.”

Fixing for system reliability whereas AI remains to be in its early days of adoption is each Saran’s ardour and the issue that retains him up at night time, he says.

“I believe making AI methods genuinely auditable and clear is among the extra vital engineering issues of the subsequent decade,” he says, “and it doesn’t get talked about as a lot as mannequin capabilities do.”

IEEE can play a job in shaping AI improvement

Saran joined IEEE in March 2025.

“After greater than 20 years of constructing knowledge infrastructure throughout organizations like Yahoo, Intuit, Meta, and Wealthfront,” he says, “I needed to contribute again to the technical neighborhood by way of peer assessment, publishing, and interesting with the broader engineering analysis ecosystem.”

He says he believes IEEE can play an essential role in shaping the way forward for AI by way of its standards work.

“We have now requirements for all the pieces, from constructing codes to software program testing,” he says. “AI infrastructure wants one thing related, and IEEE has the credibility and the attain to try this.”

One other downside that IEEE can resolve for, he says, is entry: “IEEE operates in loads of nations the place engineers don’t have straightforward pathways into the worldwide skilled neighborhood. That form of connection issues, particularly as AI improvement spreads past a handful of organizations in a number of cities.”