In December, we’re launching work on an automated code review system based on generative artificial intelligence. The project aims to support internal code quality processes using cutting-edge AI technologies. The system will analyze code from multiple perspectives: including vulnerabilities, compliance with good coding principles, performance anti-patterns, and bad code smells. The solution will evolve from using public APIs (OpenAI, Anthropic), through our own cloud infrastructure, to (possibly) an on-premises solution. We’re also considering proprietary solutions based on our own datasets and quality metrics. We believe that integration with CI/CD pipelines and feedback loops in Pull Requests, combined with these metrics, will enable data-driven continuous improvement of development processes in our company.
