AI code testing startup CodiumAI told QA Financial it has launched a new AI-powered, real-time error testing tool.
CodiumAI, which runs a generative artificial intelligence code testing platform for developers, said it has finalised the development and testing of its new ‘proactive AI agent’ product.
The solution is able to generate codes to fix identified errors or flawed codes the tool has picked up before the code generation process has been completed, explained CodiumAI co-founder and chief executive officer Itamar Friedman.
He pointed out that the new platform is an expansion of the product CodiumAI introduced a year ago, when the Israeli startup raised $11 million and launched CodiumAI Coding.
A range of large banks and financial institutions in Israel and across the U.S. and Canada went on to use and integrate the platform into their digital infrastructure. Friedman said his firm plans to target the same client base this time round.
“There are a lot of concerns about the quality of AI-generated code.”
– Itamar Friedman
Powered by OpenAI’s GPT-4, the platform supports developers to test various codes in order to write the most efficient and flawless ones.
With AI Agent, Friedman said it is able to not just test codes but also identify and fix codes that are deemed incorrect, flawed or weak.
AI Agent, which is supported by code completion tool GitHub’s Copilot, places a particular emphasis on the quality of the AI-generates code structures, Friedman explained.
He said developers are increasingly looking for AI tools that can assist them with the tedious tasks that are essential for ensuring high quality code and correctness.
“We are looking to an ‘agentic future’ where busy developers will team up with agents to improve code quality and processes,” Friedman pointed out.
“Integrated agents will be there every step of the way, offering insights, recommendations and solutions proactively,” he continued.
Real-time testing
Code testing is vital in any software development process. The traditional approach is that testing only takes place after codes have been generated.
Friedman emphasized his company’s AI Agent solution is able to test codes while the developer is still working on.
He stressed there is a clear appetite for its new product as more than nine in 10 developers now use generative AI tools to help them write codes.
“But although these tools can help humans to write code faster, there are a lot of concerns about the quality of AI-generated code. That is what we intend to fix,” Friedman concluded.
THIS COMING THURSDAY
Stay up to date and receive our news, features and interviews for free
Our e-newsletter lands in your inbox every Friday. Sign up HERE in one simple step.
READ MORE
