Vineet Edupuganti and the work after the warning

A battery is usually judged by how long it lasts. Vineet Edupuganti built one whose eventual disappearance was part of the point. As a high-school student in Portland, he worked on a biodegradable power source, designed to dissolve after its useful life. In December 2016, the project earned him the $100,000 individual prize in the Siemens Competition. It was a substantial reward for asking an unusual engineering question: what should happen when the useful work is over?

A decade later, Edupuganti is co-founder and CEO of Cogent, working on a problem with a less cooperative ending. A security tool finds a vulnerability. A warning appears. Someone must investigate it, locate the person responsible, arrange the repair and confirm that it worked. The warning can be generated quickly. The rest can become everybody’s problem, which is an excellent way of making it nobody’s next task.

His career has moved from materials to machine learning to enterprise software. Read across those changes and a practical interest emerges: the work required to carry a technical idea into use. The connection is an interpretation of his path, rather than a claim that a teenage experiment contained a startup waiting to hatch.

A battery with an exit plan

At Oregon Episcopal School, Edupuganti investigated transient electronics, devices designed for a limited useful life. His battery project had possible applications in environmental sensing, including equipment that could do its job without leaving the same burden of discarded hardware. Durability, the ordinary virtue of electronics, became a design decision rather than an unquestioned goal.

The materials mattered. By 2017, he described extending battery life sixfold through a change from magnesium to a magnesium alloy. Better resistance to corrosion allowed the battery to operate longer before its planned disappearance. The objective required a balance: something had to last long enough to be useful, then stop being a disposal problem.

At the national Siemens finals, hosted at George Washington University, he won the top individual award. In 2017, he was also named one of the Regeneron Science Talent Search finalists. These were distinct recognitions, attached to a period when his work still belonged to the science exhibition rather than the enterprise sales meeting.

He was already thinking about what followed the exhibition. While still in high school, he wanted to find practical uses for research and eventually bring them to market as an entrepreneur. The ambition predates Cogent by years. It also makes the subsequent move toward product work easier to understand.

The point guard changes subjects

His first career plan involved a different sort of performance. As a child, Edupuganti wanted to become a professional basketball player. The attraction included the numbers as well as the game. Keeping track of statistics sounded like a pleasing addition to being paid to play.

His later account of that plan had some welcome self-mockery. Being tall in his grade, he discovered, offered limited assurance of a professional future. Interests shifted. A seventh-grade attempt to make solar cells with a kit and fruit juices failed, but left him curious about materials and their interactions.

“My experiment didn’t quite work out.”

Vineet Edupuganti, recalling an early solar-cell project, 2017

There were other interests alongside the research. In high school, his interests included classical piano, guitar and varsity tennis, alongside tutoring math and reading in Spanish. He was a Portland Trail Blazers fan. The picture is pleasantly crowded: a student with experiments to run, music to practise and a hometown team capable of consuming whatever attention remained.

The sports interest survived the change in professional plans. At Stanford, he played basketball, tennis and golf in his spare time, admitting that watching all three went better than playing them. It is a small detail, but a useful interruption to the usual sequence of prizes and job titles.

Stanford, then the machinery behind the product

Edupuganti studied electrical engineering at Stanford, graduating with his bachelor’s degree in 2021. The university’s 2022 commencement program also lists him for a master’s degree in computer science. The combination places his education across both the physical and computational sides of engineering.

Before Cogent, his experience included Facebook and New Relic. He worked in engineering and research roles at both before a summer internship at Abnormal Security. It was a useful intermediate moment: a student with an interest in entrepreneurship, approaching the company where he would build much of his operating experience.

At Abnormal, he worked as a machine learning engineer on threat detection before moving into product management. His responsibilities included infrastructure and platform initiatives, integrations and data loss prevention. He later led the Multi-Product Platform Division.

Those roles concern how a product fits into the rest of an organization. Integrations connect it to other systems. Platform work creates foundations that several products can share. Moving between engineering and product leadership means considering both what the system can do and what a customer needs to accomplish with it. That background helps explain why the operational details receive so much attention in his next company.

Three founders, one unfinished job

Cogent was founded in 2024 by Edupuganti, Geng Sng and Thanos Baskous. Edupuganti serves as CEO, Sng as chief technology officer and Baskous as vice president of engineering. Sng also came from Abnormal; Baskous brought experience from Coinbase and Blackstone.

The company publicly launched in July 2025 with $11 million in funding. Its chosen territory was vulnerability management, with particular attention to the tasks between identifying a weakness and getting it resolved. That is an awkward stretch of work because a finding can cross technical systems and organizational boundaries.

Cogent describes AI agents that investigate findings, establish ownership, evaluate the risk in context, develop remediation steps and track the result. The proposed value sits in the sequence. Each task supplies something the next task needs. An answer without an owner is difficult to act on; a change without verification leaves another question open.

Five principles, very little patience for friction

Edupuganti has put the product philosophy into five principles: simplicity, completeness, autonomy, self-service and a product-first approach. The list is short. Its consequences for the builder are less modest. A product that promises to cover an entire workflow has fewer opportunities to hand the troublesome bits back to its customer.

Self-service, in this formulation, means that users should be able to handle configuration and integrations directly. Autonomy means agents doing work with awareness of business constraints. Simplicity has to coexist with both. There is considerable engineering hidden inside the apparently innocent promise that something should be easy to use.

The principles also make friction a product question. If customers repeatedly struggle at the same point, the company’s stated response is to change the software. That assigns responsibility to the people building the tool. It is a useful standard to publish, because customers can remember it.

Trust enters the same conversation. In a Greylock discussion with Abnormal CEO Evan Reiser, Edupuganti explored agentic security, explainability and the relationship between people and automated work. For a security product, understanding an action is part of deciding whether to allow it. A helpful assistant still needs to account for what it has done.

The money buys another set of questions

In February 2026, Cogent announced a $42 million Series A led by Bain Capital Ventures, with continued backing from Greylock Partners and Definition. The round brought the company’s disclosed funding to $53 million. Enrique Salem joined the board. Edupuganti described plans to invest in the data and AI platforms, commercial operations and hiring.

The figures describe resources available to a company. They carry no automatic verdict on how well an individual customer’s security program will work. The important test remains closer to the original problem: whether a finding gets investigated, assigned and resolved with less delay and confusion.

That distinction keeps a founder profile in proportion. A science award recognizes a research project. A degree records study. A funding round enables another stage of company building. Edupuganti has accumulated all three kinds of milestone, but each answers a different question.

Back to the beginning, with a different problem

By October 2026, his product writing had moved into attack-path analysis. Cogent’s announcement described connecting information about assets, identities and controls to examine routes an attacker could take. The emphasis extended beyond an isolated weakness to the chain of conditions that might make it consequential.

The company described agents working backward from important assets, testing possible routes against evidence and recommending where a change could break a path. The explanation reflects a continuing interest in context: which pieces matter together, and which intervention would change the outcome?

That month, Stanford published his Entrepreneurial Thought Leaders conversation with lecturer Ravi Belani under the title First Principles in the AI Era. The former undergraduate had returned as a founder discussing how to reason about building in a changing technological environment.

The battery and the security company remain different undertakings. Still, both make the ending part of the engineering problem. A battery needs to serve its purpose and then disappear as intended. A security finding needs to reach a verified resolution. Edupuganti’s current work lives in that final stretch, where an interesting technical capability must become something a person can actually use.