AI engineers need to carry out complex tasks by understanding the user perspective and context of the problem. They must know the location of the bug, its nature, impact on the product, downstream changes, and more before taking action. Context is gathered from sources like Jira tickets and code base. Writing detailed specs and code planning is crucial for AI engineers. To complete tasks effectively, AI agents must understand the entire process end-to-end. Some companies in this space include Devin, Factory, Code Gen, and Suite Agent. Having a code-specific model is believed to be essential for long-term differentiation in the code app layer.