I Built an AI Agent That Optimizes Code — But Benchmarks Decide What Survives

说明与用途

一个自治代理提出代码优化，并通过冻结的基准测试和单元测试来客观评估其是否真正提升性能和正确性。

问题

AI coding agents can propose optimizations that look good but may not actually improve performance or correctness.

结果

The system automates performance optimization by letting an AI agent iterate through changes, with a benchmark and test suite acting as an external judge to ensure only genuinely faster and correct code is retained.

工作流

• Agent proposes an optimization
• Agent applies the change
• Run unit tests
• Run frozen benchmark
• Measure performance
• If performance improves and tests pass, KEEP the change; otherwise DISCARD
• Repeat

证据

The agent does not decide whether its own change is good. Instead, every experiment is evaluated by a frozen benchmark and test harness.

来源

If the change improves the measured result and still passes the required checks, it can survive.

来源