How Ramp engineers accelerate code review with Codex
Quick Answer
Ramp engineers leverage Codex with GPT-5.5 to streamline code reviews, achieving substantial feedback in minutes rather than hours.
Quick Take
This integration significantly enhances productivity and accelerates the deployment of improvements.
Key Points
- Codex and GPT-5.5 reduce code review time from hours to minutes.
- Engineers receive substantive feedback quickly, enhancing workflow efficiency.
- The integration allows for faster deployment of code improvements.
- Ramp's approach exemplifies the benefits of AI in software development.
📖 Reader Mode
~4 min readAt Ramp, engineers are using Codex with GPT‑5.5 to accelerate code review and develop internal agentic tooling, helping teams get substantive pull request feedback in minutes instead of hours. Thanks to its reasoning capabilities, Codex with GPT‑5.5 is uniquely able to reduce the amount of manual, hands-on work they’d otherwise have to do.
“Codex code review catches things that I miss and that other engineers miss and that other AI code reviewers definitely miss.”
—Austin Ray, AI DevEx at Ramp
Running code reviews that teams can rely on
Ramp’s AI Developer Experience team is using Codex to improve software development velocity and code quality.
“Codex code review is industry gold standard. We’ve been relying on it for a long time here at Ramp,” explains Austin Ray, who leads AI DevEx. “It’s incredible, and our engineers ask for it by name. They look forward to its comments on every PR, and it’s become a mandatory part of a lot of code review flows.”
Ramp engineers who used to wait hours for a first review can now get substantive feedback from Codex in minutes. Codex stands apart from other tools because it deeply reasons against the codebase, resulting in what Ray describes as “a level of thoroughness that most human reviewers don’t have time for.”
Codex matches this depth with an experience that, Ray says, “meets engineers where they are.” Engineers who prefer to work close to the metal can work from the CLI, and the Codex app provides visual cues, utilities, and additional features for those who want them. Ray, typically a CLI user, felt drawn to the app. “It feels like the app shepherds you toward higher productivity in your engineering workflows,” Ray says.
“Codex with GPT-5.5 is incredibly adept at dealing with that complexity in a way that would take me a ton of mental effort, a lot of sleep, and a lot of single-minded focus on the problem to figure out.”
—Austin Ray, AI DevEx at Ramp
Developing internal tools with Codex
Ray is also using Codex to support the development of On-Call Assistant, an agentic tool that takes on most of the burden for Ramp engineers during on-call rotations.
“On call is hard,” Ray explains. “We have a lot of business logic, domain knowledge, and heavy incidents. You have to keep a ton of things in context and reason through a lot of complexity.”
For an engineer, this can be difficult. It takes a lot of mental effort and even more single-minded, unbroken focus.
“There’s just a ton of complexity,” Ray says. “There are plenty of concurrency bugs, a tricky balance to strike between external events and internal events, and long-running incident investigations with evolving details you have to keep working in.”
With Codex, Ray can depend on its “incredibly adept” reasoning capabilities to support development. As a result, On-Call Assistant has become significantly faster to build, and Ray is more confident about every improvement shipped.
“Our product surface area is pretty immense,” Ray says. “Codex with GPT‑5.5 handles it like it’s nothing.”
Leadership lessons
Ray is a platform engineer, first and foremost, and he evaluates all developer tools, including AI-driven ones, through that lens. As he puts it: “Does it actually change how people ship code, or is it just a demo?”
And that’s what Ray recommends for other leaders: Focus on the hands-on experience and real-world results.
- Demonstrate the potential of AI tools first-hand: “Get your engineers to install Codex, sit down with them, and guide them through a really solid first session. Paint the picture of what development could be for them.”
- Build a path to trust and iteration: “Most engineers don’t fully understand or trust that they’re going to have a good experience with this. They treat it as something experimental. By guiding them through that first experience, you change their perspective and make them willing to explore and iterate themselves until they become one of your best AI users.”
- Invest in the feedback loop: “We work directly with the Codex team on feedback. When we hit issues, we have a direct line. That feedback loop is what makes a vendor relationship worth investing in, and we’ve made incredible progress with the Codex team.”
“Codex is the real deal. Codex definitely helps us ship faster.”
—Austin Ray, AI DevEx at Ramp
What’s next
Codex is changing how fast Ramp engineers can work and giving them the resources to support even greater ambitions. For Ray, this indicates a new way to think of engineering as a whole.
“Engineers are going to become orchestrators. The skill is no longer writing every line of code yourself. It’s knowing how to direct AI tools like Codex, when to trust them, and when to push back. At Ramp, our best engineers learn that fastest.”
— Originally published at openai.com
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from OpenAI Blog
See more →Scientific computing in the age of agentic AI
AI agents are transforming scientific computing by streamlining software development, enabling researchers to focus on discovery. Projects using Codex and Claude Code report accelerated development and improved maintenance, though challenges in validating AI outputs remain. Long-term stewardship of research software is crucial to ensure reliability and reproducibility.