Playerzero raises $15 million to prevent AI agents from shipping buggy codes

New problems are created as Silicon Valley is running towards a future where AI agents do most of their software programming. Finding the generated bug before producing AI-generated bugs. The same goes for opening To address such problems, A former employee explained it.

Newly funded startups playerzero I created a solution. Using an AI agent trained to find and fix problems before code is produced, startup CEO and proprietary founder Animesh Koratana tells TechCrunch.

Koratana created PlayerZero while he was at Stanford Dawn Lab for machine learning under Matei Zaharia, an advisor and lab founder. Zaharia is, of course, a well-known developer and co-founder of Databricks. He created the basics of that while working on his PhD.

Playerzero announced on Wednesday that the $15 million Series A, led by Foundation Capital’s Ashu Garg, is an early Databricks backer. This follows $5 million seed led by Green Bay Ventures and several notable angels, including Zaharia, Dropbox CEO Drew Houston, Figma CEO Dylanfield and Versel CEO Guiller Morauch.

During his stay at Stanford Dawn in Koratana, now 26, he is working on AI model compression technology and “was exposed to language models really early on,” he says. He met the developer who created some of the first AI coding aid tools.

He then tells TechCrunch, “There’s this world where computers write code. It’s not going to be human anymore.” “What does the world look like at that point?”

He knew the term “AI Slop” before it was even coined to write code that would break things the same way these agents would oversee human supervision.

That problem is also exacerbated by so many agents who are writing far more code than ever before. It is not always practical for humans to check all AI writing codes for bugs and hallucinations. And this problem becomes even more intense for the large, complex codebases that companies rely on.

Playerzero trains the models with a “very deep understanding of the codebase, and an understanding of the way they are built, the way they are built.”

His technology studies the history of corporate bugs, problems and solutions. When something breaks, his product “can get the reason and fix it and learn from those mistakes to ensure that it doesn’t happen again from those mistakes,” says Koratana. He likens the product to a large code-based immune system.

Landing his advisor, Zaharia, as an angel, was the first step into the fundraiser, but the moment he truly validated his idea was when he showed another well-known developer, Rauch, a demo. Rauch is the founder Triple-Unicorn Developer Tool Company Versel and creator of the popular open source JavaScript framework next.js.

Rauch watched the Koratana demonstration with interest, but saw skepticism and asked how much “real” it was. Koratana replied that this is the code “running in production.” This is a real example. He was quiet,” Koratana said. Then his immediate investor replied, “If you can actually solve this in the way you imagine it, that’s really a big deal.”

Of course, PlayerZero is not just trying to solve bug issues generated by AI. Last week, Anysphere cursor I launched a bugbot One example is to detect coding errors.

Still, PlayerZero has already gained traction to focus on a large codebase. It was thought for the world where agents are coders, but is now used by several large companies using coding coding coding facilities. For example, Zuora, a subscription billing company, is one of the startup marquee customers. Zuora uses technology across the engineering team, including monitoring the billing system, the most valuable code.