AI-Assisted Software Development
Practical use of OpenAI Codex and Claude Code for investigation, implementation, review, and modernization with human verification.
Discuss your project ↗Use AI tools as engineering collaborators
OpenAI Codex and Claude Code can accelerate codebase exploration, repetitive implementation, test creation, documentation, and review. They are most effective when the task has a clear outcome, the relevant context is available, and every change is verified against the actual system.
I use AI development tools alongside Visual Studio Code and Visual Studio 2026 while keeping engineering judgment responsible for architecture, security, correctness, and release decisions. A useful workflow asks the tool to gather evidence, make a bounded change, and validate it rather than accepting a large unreviewed rewrite.
AI-assisted workflows
- Understand an unfamiliar repository and trace behavior.
- Implement focused changes with tests and verification.
- Review code for regressions, edge cases, and maintainability.
- Modernize repetitive or legacy code incrementally.
- Create documentation that stays close to the implementation.