Itamar
Friedman

In March 2026, Qodo closed a $70 million Series B led by Qumra Capital, pushing the company's total funding to roughly $120 million. The round drew in enterprise-software specialists and a pair of notable angels: Peter Welinder of OpenAI and Clara Shih of Meta. For a company founded in 2022, it was a fast climb - and a clear signal that the market was starting to worry about the same thing Friedman had been warning about for years.

The worry is simple to state. AI assistants now write a large and growing share of the code inside modern software. Someone still has to confirm that the code is correct, secure, and consistent with how a given company already builds things. Friedman's argument is that the tools built to generate code are the wrong tools to judge it. "Generating systems and verifying systems require very different approaches," he says. "Code generation companies are largely built around LLMs. But for code quality and governance, LLMs alone aren't enough."

That distinction is the whole company. Qodo builds AI agents that review, test, and govern code rather than produce it. Its multi-agent review system, Qodo 2.0, learns an organization's standards and past decisions, then applies them to new code. It ranked first on Martian's Code Review Bench with a 64.3% score, and it now runs inside enterprises including Nvidia, Walmart, Red Hat, Intuit, Texas Instruments and Monday.com.

Friedman keeps returning to a word that sounds almost philosophical for a developer-tools company: context. "Quality is subjective," he says. "It depends on organizational standards, past decisions, and tribal knowledge. An LLM can't fully understand that context." The pitch to a bank or a retailer is not that Qodo writes better code. It is that Qodo remembers what "good" means at that specific company, and holds every change to it.

He frames the moment as a shift in what AI is for. "Every year has had a defining moment - from Copilot to ChatGPT to full task automation," he told TechCrunch. "Now we're entering a new phase: moving from stateless AI to stateful systems - from intelligence to 'artificial wisdom.'" Intelligence generates. Wisdom knows whether it should. Qodo, in his telling, is infrastructure for the second thing.

Friedman's origin story has a small joke buried in it. As a teenager in Israel, a teacher asked the class something about "Windows." He looked at the actual windows in the room. The confusion, as he tells it, kicked off a fascination with computers that led him to build his high school's network and start writing web applications in the 1990s.

He studied at the Technion - Israel Institute of Technology, earning a BSc and then an MSc in Electrical Engineering, summa cum laude, focused on machine learning and computer vision. Early on he worked at chipmaker Mellanox on design and verification - an unglamorous discipline about making sure complex systems actually do what they claim, a theme that would resurface two decades later.

In 2012 he co-founded Visualead as CTO. The company built machine-vision technology and visual QR codes for mobile devices, and it grew fastest in Asia. In 2016, Alibaba Group acquired it. Friedman stayed on for about four and a half years, eventually serving as Director of Machine Vision and head of Alibaba's machine vision lab in Israel, working within the company's DAMO Academy on foundation models.

He is candid that the exit felt incomplete. The returns were good, but the acquisition was a detour from the original ambition. That restlessness is what pushed him back to zero in 2022 to co-found CodiumAI, later renamed Qodo. His advice to founders who try to do everything at once comes from lived experience juggling a master's degree and a startup: "Choose one and run with it fast, deep."