Unblocked AI Code Review: how it works, and why we built it with a founding engineer

Richie, a founding engineer at Unblocked, discusses our feature-rich code review product, emphasizing the importance of context in AI-driven code reviews. He addresses common problems in traditional code reviews, such as latency, human biases, and inefficiencies, and explains how Unblocked's system utilizes context from various sources like Slack conversations, documentation, and previous reviews to solve these issues. Richie also highlights their learnings from the CI failure agent and how these insights shaped the development of their AI code review tool, aiming to find genuine bugs and improve the efficiency of merging pull requests. The conversation explores the challenges and methodologies in designing an effective AI code review system, including the use of adversarial models and human-centric design principles.

We also have a standalone blog post covering this episode here:

https://getunblocked.com/blog/building-ai-code-review-a-conversation-with-richie/