الباحثون

Tanu Malik

المنشورات 2

نسخة أولية وصول مفتوح

Code That Works, Environments That Don't: Measuring Environment Reproducibility in AI-Generated Software

Code generation has emerged as a central capability of large language models, with coding agents now able to produce functionally correct software projects from natural language prompts. However, functional correctness alone does not capture a critical dimension of generation quality: environment specification, defined …

المؤلفون المشاركون