When a candidate submits their code for evaluation, it will undergo a Code Review performed by an AI. This review uses the Code Review Scorecard defined in the corresponding coding test challenge.
The code review criteria can be customized to meet your hiring needs. Here are some tips to improve the AI review process:
💡 To edit the scorecard of one of Alva’s default challenges, you must first duplicate it!
Be specific. Clearly defined instructions lead to better reviews. For example, instead of asking the AI to comment on the database structure, provide specific pointers such as, “What datatypes have been used for the database structure?”
Break down complex requirements into multiple criteria. Instead of having one criterion with a long list of instructions, breaking it down into specific dimensions can help the model conduct a more thorough review.
Provide context. Ensure the AI understands the purpose and context of the code. For example, mention the primary functionality the code is supposed to achieve or specific constraints it must adhere to. This will help the AI provide more relevant and insightful feedback.