Snowflake / Engineering quality
Code review, with AI and human judgment
Agentic code-review skills that evaluate correctness, performance, security, and test quality while keeping humans in the loop.
The challenge
Scale code review without losing the judgment and accountability that make engineering reviews valuable.
My contribution
Ideated and built AI code-review skills across correctness, performance, security, and test-case quality, combining automated evaluation with human-in-the-loop review.
The outcome
Delivered more than 500 PR reviews in one month, reduced review cycle time by 50%, and strengthened the reliability of AI adoption in the team.