Shwetha Devanga

Evidence

The same discipline, applied at work

At Veracode I led product on an AI-powered remediation engine — concept to enterprise deployment on AWS Bedrock — and defined its evaluation system: fix correctness, regression safety, false-positive rate.

10x

Improvement in AI fix success rate, launch baseline to steady state

11x

Growth in monthly active users of AI-assisted remediation

5x

Growth in enterprise accounts with sustained active usage

~30%

ML experimentation operating cost cut via re-architecture

9+

Teams aligned across product, ML, security research, and platform

96%

Committed story points delivered, sustained across two Scrum teams

Relative figures from internal product dashboards and Jira, Apr 2024 – Jul 2026. Full context and methodology available on request.