In a significant development for the software development industry, a New York-based startup named Qodo has successfully raised $70 million in a Series B funding round led by Qumra Capital. The company, co-founded by former Alibaba executive Itamar Friedman, aims to empower organizations to produce reliable, secure code that adheres to both internal and external standards. Qodo’s mission is to help businesses mitigate potential risks, which include system outages, security vulnerabilities, and failures to comply with regulatory requirements. Friedman emphasized that the startup focuses on navigating and addressing critical aspects of software development that often remain hidden beneath the surface, such as maintainability, security, compliance, and review processes.

In remarks to ISMG, Friedman elaborated on the pressures of modern software development, stating that developers must grapple with numerous underlying factors that can affect outcomes. “When you’re talking about real-world software development, you have to deal with everything underneath the glacier, under the water, under the surface,” he explained. “All that is super critical.” Founded in 2022, Qodo has grown rapidly, employing 123 individuals and raising a cumulative $120 million, reflecting increasing investment interest in the sector. Prior to launching Qodo, Friedman held significant roles at Alibaba, where he spent four years honing his skills in machine vision and other technological advancements.

Governance in Code Generation: A Growing Necessity

Friedman pointed out that while AI systems can generate thousands of lines of code in a matter of minutes, this rapid development cycle also introduces new vulnerabilities. The absence of robust governance mechanisms often means that faster code delivery leads to quicker failures. He highlighted the necessity for governance tools that ensure generated code aligns with architectural standards—this includes meeting performance expectations and respecting organizational policies and ethical values.

“It’s becoming increasingly clear that code generation alone is not sufficient,” Friedman remarked. “LLMs are enablers. They’re the ‘why now,’ but they don’t encompass the entirety of the solution.” The challenges posed by AI-generated code stress the importance of verification and alignment with set expectations. This endeavor necessitates advanced AI models and sophisticated methodologies to continuously analyze code across entire organizations.

Friedman illustrated this point with an example of how existing tools function, questioning their effectiveness. “Have you ever seen a code generation tool like Claude Code tell you, ‘Oh, sorry, I can’t do this for you’? All it tries to do is please the developer,” he said. This tendency calls for systems that not only generate code but also verify its quality and compliance. Qodo, therefore, aims to operate with a swarm of specialized agents that evaluate every code alteration. This innovative multi-agent system approaches quality assurance from various angles, ensuring that code is assessed for correctness, security vulnerabilities, and adherence to best practices.

Enhancing Code Quality with Tailored Solutions

Recognizing the subjective nature of code quality, Friedman stated that standards can differ significantly among organizations depending on their risk tolerance and operational requirements. Qodo’s approach incorporates internal data—such as developer discussions and existing codebases—to customize evaluations to the context in which the code is developed. “Qodo is building the second brain,” he said, introducing a system of record that captures vital information and insights to facilitate better decision-making throughout the software development lifecycle.

Traditional coding agents generally operate without retaining knowledge of prior interactions, leading to inconsistencies in decision-making. To address this shortcoming, Qodo aims to build a persistent memory layer that archives essential patterns and rules over time, enhancing the system’s decision-making capabilities. Friedman remarked, “We released the most advanced multi-agent system, reaching top performance on code review benchmarks efficiently.” He further mentioned that Qodo is pioneering efforts to recast quality as something inherently subjective, thereby shifting industry perceptions.

The general expectation from coding agents is to deliver results based on immediate requirements. However, the emphasis of Qodo is to introduce necessary checks that balance out this bias. Rather than replacing LLMs, Qodo intends to function alongside them, enforcing rules and standards to safeguard quality outcomes. By identifying deviations from best practices or potential risks, Qodo significantly enhances the development process.

In closing, Friedman stated the importance of the review process, saying, “You need to write the code, and then you do review and testing. And right now, what’s happening is this is shrinking, shrinking, shrinking.” As organizations continue to navigate the complex landscape of software development, Qodo’s innovative approach to governance presents a crucial path forward for ensuring secure, efficient code generation.