← All organizations

Code quality, security, and AI code verification

Sonar

Sonar builds tools that help developers find quality and security issues in human-written, AI-generated, and third-party code before production. Its SonarQube family combines static analysis, automated code review, and remediation across editors and CI/CD pipelines. Teams can choose managed SonarQube Cloud or self-managed SonarQube Server, while SonarQube for IDE provides feedback as developers write code. Customizable quality gates enforce standards before merging or release, and AI CodeFix uses language models to suggest repairs.

Olivier Gaudin, Freddy Mallet, and Simon Brandhof started the open-source project behind Sonar; the company was incorporated in 2008. Gaudin is now chairman, and Tariq Shaukat is CEO. Its engineering evolved from integrating existing analysis tools to building its own symbolic execution, semantic analysis, and dataflow analysis. Sonar also introduced Cognitive Complexity in 2016: a metric that estimates how difficult control flow is to understand, accounting for nesting and interruptions while ignoring readable shorthand.

In 2026, Sonar reported that 7 million developers and more than 75% of Fortune 100 companies relied on SonarQube. It also acquired Gitar, an AI code review and validation platform that fixes code, bugs, and CI failures within GitHub and GitLab pull requests. The acquisition expanded Sonar’s product scope, with Gitar remaining available separately and integration with SonarQube planned.

www.sonarsource.com

Start here

  1. Can LLMs generate Enterprise Quality Code? — Prasenjit Sarkar, Sonar

    Start here to learn how code volume and security findings provide different lenses for comparing coding models.

    Prasenjit SarkarAI Engineer Europe 2026

  2. Guide, Verify, Solve: The Engineering Discipline Agentic Development Demands

    Choose this talk for a guide-verify-solve approach combining agent guardrails, automated checks, and iterative remediation.

    Anirban ChatterjeeAI Engineer World's Fair 2026

  3. In the Land of AI Agents, the Verifiers Are King

    Read this session for the lifecycle perspective: how verification, CI, and maintenance fit together as agent capabilities expand.

    Tariq ShaukatAI Engineer World's Fair 2026

Messages from the stage

Quality beyond benchmark success

Prasenjit Sarkar distinguishes functional benchmark performance from security and maintainability, using SonarQube analysis of Java assignments to compare model outputs.

Human acceptance is not verification

Anirban Chatterjee discusses research on acceptance of incorrect AI advice alongside reported increases in static-analysis warnings and code complexity, motivating guardrails and automated verification.

Maintenance inside the agent cycle

Tariq Shaukat presents architectural context, layered quality gates, and technical-debt remediation as connected parts of the Agent Centric Development Cycle.

Affiliations reflect each recorded session, not necessarily current employment.

Company sources · checked 2026-08-28