Preparing for technical interviews can feel like a solo sport, especially for students without access to mentors, mock interviews, or insider advice. In this practice-based report, we explore how publicly available LLMs (e.g., ChatGPT, GitHub Copilot) can be harnessed via zero-shot prompting, not as a shortcut to answers, but as collaborative scaffolds that reinforce genuine skill development. Building on an equity-focused framework for interview preparation, we introduce five responsible prompting patterns that guide students to: (1) restate and validate problems, (2) generate clarifying questions, (3) brainstorm alternative solutions, (4) verify Big-O analyses, and (5) perform self-code reviews without an IDE. We argue that, when used responsibly, LLMs have the potential to amplify student agency, reduce interview anxiety, and democratize access to high-quality preparation.