Sonar, a leading provider of code quality and code security solutions, has acquired AutoCodeRover. AutoCodeRover was founded by researchers from the National University of Singapore to help enterprises and developers maintain reliable and performant software systems through autonomous program improvement.
AutoCodeRover was co-founded by Ridwan Shariffdeen, CEO; Martin Mirchev, CTO; Yuntong Zhang, Co-Chief Scientific Officer; Haifeng Ruan, Co-Chief Scientific Officer and Abhik Roychoudhury Provost’s Chair Professor of Computer Science at the National University of Singapore and advisor to AutoCodeRover.
The acquisition of the autonomous AI agent platform for software development strengthens the Sonar platform. In a single acquisition it delivers a strong platform for the development of its AI capabilities for the code quality and code security solutions platform.
Tariq Shaukat, CEO of Sonar, said, “AI agents provide developers and companies with a powerful new set of tools to build better, faster. Studies estimate that developers spend almost half their time fixing bugs, addressing technical debt, and on other work that doesn’t help create the future. This leads to lower developer productivity and happiness. Agentic AI working side by side with developers will help them build better, faster, and spend more of their time on the activities they enjoy – namely, writing code, creating new products, innovating, and driving impact.”
The deal
Sonar did not disclose the terms of the deal. However, Sonar will continue to collaborate with researchers from the Trustworthy and Secure Software Research Group. Out of which AutoCodeRover was spun out. Professor Abhik Roychoudhury leads that Group in the ongoing improvement of agentic AI technology.
Prof. Tan Kian Lee, Dean NUS School of Computing commented, “We are proud to see Professor Abhik Roychoudhury’s spinoff, AutoCodeRover, acquired by Sonar. This highlights the innovation and the real-world impact of our research. We look forward to seeing how this success continues to shape the future of software development.”
Enterprise Times also asked Harry Wang, VP of Growth & New Ventures at Sonar, about who would be joining Sonar under the deal. He replied, “All founding members will be joining Sonar in different capacities. Adbhik Roychoudhury will remain to be a full-time professor at the National University of Singapore (NUS) and serve as an advisor to Sonar.”
As of writing, none of the co-founders had updated their LinkedIn profiles with their new roles. Sonar did not reveal how many of the AutoCodeRover team will join the company. However, Wang noted, “We’re very excited about this not just being an acquisition of technology, but the beginning of Sonar’s commitment to investing in R&D in Singapore.”
Ridwan Shariffdeen, CEO and Co-Founder of AutoCodeRover, commented, “We are thrilled to join forces with Sonar and contribute to the mission of helping developers around the world accelerate their work. By combining our advanced agent technology with Sonar’s industry-leading code quality and security solutions, we’ll drive even greater impact for developers and organizations worldwide.”
What is AutoCodeRover
The work the team has done on its nascent product has led to two recent academic papers. These include:
- AutoCodeRover: Autonomous Program Improvement (ISSTA24)
- SpecRover: Code Intent Extraction via LLMs (ICSE25)
AutoCodeRover leverages several LLMs to examine code and propose code modifications and improvements within the context of the whole program. The team has built unique code and reasoning alongside automation capabilities that can make a huge difference in code quality.
It infers the programmer’s intent by leveraging artefacts within each program. It builds natural language descriptions of each function. Then builds multiple alternate lines of code and assesses them to identify the optimal solution. The complete solution automates much of the Software Development Lifecycle (SDLC).
An SWE-bench evaluation recently demonstrated proof of its capabilities. SWE-bench aims to score solutions on whether language models can resolve real-world GitHub issues. It scored as follows:
- SWEBench Verified 46%, placing it 23rd
- SWE Bench Lite 37%, placing it 17th
- SWEBEnch Full 25%, placing it 2nd highest!
NB: Placings accurate as of writing the article.
The team has built connections to several LLMs, including OpenAI, Anthropic, Gemini, AWS Bedrock, Llama3 and Azure AI. The code already supports eight mainstream languages and has plans to add others. The languages currently include:
- C
- C++
- Go
- Java
- JS
- PHP
- Python
- TS
Looking forward
The key part of this acquisition is how the technologies from the two companies will combine. Sonar intends to integrate AutoCodeRover with SonarQube. This is a tool that continuously inspects code quality to perform automatic reviews with static analysis of code to detect bugs and code in over 30 different languages.
The combined solution will help developers in the following ways, according to Sonar:
- Accelerate Development: Automate key tasks in the SDLC to reduce development time.
- Continuously Improve Code Quality: Expedite code reviews to not only detect issues but also remediate them instantly for agile delivery of high-quality applications.
- Reduce Costs: Lower development costs by handling the laborious part of coding work autonomously.
- Foster Innovation: Enable developers to spend less time on issues and more time on creating innovative business solutions.
Enterprise Times: What does this mean?
Like many acquisitions, this has huge potential for all concerned. Sonar gets access to a leasing AI-based code quality system founded by a leading research team. It also gains access to a new development function in Singapore. For the founders, it provides an early exit and the ability to focus on the tech rather than growing the business.
I asked Wang what it meant for the developers. His response went beyond that and gave an insight into how Sonar expects AI to change developer productivity. He said, “AutoCodeRover was an early-stage pre-revenue startup. We will communicate to AutoCodeRover users as we integrate AutoCodeRover into our SonarQube offering. Currently leveraged by over seven million developers and 400,000 organizations.
“AutoCodeRover enhances our solution with a sophisticated foundation for automatically addressing real-world engineering issues such as debugging, issue remediation, and code refactoring. If you look at what we’ve announced already, with AI CodeFix and AI Code Assurance, you’ll see that AutoCodeRover is essentially an extension — we suggest fixes to issues, for example, and now we have agents to do more complete tasks end to end.
“While some in the industry are saying agents will displace developers, we believe it’s more about collaboration with agents and what it means to be an engineer or developer moving up the stack to add more value. Developers and agents working together is the future, and we are committed to helping developers build better and faster with actionable code intelligence.”
