北辰青年四周 AI 实践Four-Week AI Practice with Polaris Youth
从认识自己、理解团队和用户,到制作招募物料,再到建立求职与面试 AI 工作流的一次完整实践。 A complete practice cycle: understanding myself, the team, and users; creating recruitment materials; and building AI workflows for job applications and interviews.
为什么做Why I made it
我希望不只是零散地向聊天机器人提问,而是学习怎样把 AI 放进真实任务:先想清楚任务怎么拆、步骤之间怎么连、结果怎么检查,再让 AI 参与,最后把有效的做法存下来,下次还能用。这四周就是围绕北辰青年这个真实项目,从认识自己和团队,一路做到招募物料和个人求职系统。 I wanted to move past asking a chatbot scattered questions and learn to put AI inside real tasks: first work out how a task breaks down, how the steps connect and how results get checked, then bring AI in, then save what worked so it's reusable. Over four weeks, around the real Polaris Youth project, I went from understanding myself and the team all the way to recruitment materials and a personal job-search system.
我做了什么What I did
我读了 34 名成员已开放的名片墙和个人说明书,把团队特点整理成文字,并参与小组网站「核心特点」模块的内容和交互构思;检索并整理了 9 条小红书定性样本;根据会议纪要和调研结论,指挥 AI 制作 8 张校园行动招募海报,逐页核对文案、角色名称、报名方式、Logo、字号和留白;做了一个结营复盘模板;再逐步把求职准备搭成一套工作流——岗位分析、经历匹配、简历核验、面试准备、面试后复盘——并加了「面试官视角」这一层,最后沉淀进个人知识库。 I read the shared profile wall and personal guides of 34 members and wrote up the team's characteristics, and contributed to the content and interaction concept for the team site's “core traits” module; I collected and organised nine qualitative Xiaohongshu samples; from meeting notes and research findings I directed AI to make eight campus-recruitment posters, checking every page for copy, role names, sign-up method, logo, type size and spacing; I built a debrief template; and I gradually assembled job-search preparation into one workflow — role analysis, experience matching, résumé fact-checking, interview prep, post-interview debrief — with an added interviewer-perspective layer, finally stored in a personal knowledge base.
AI 帮了什么How AI helped
AI 承担了资料整理、文案与视觉草案、页面代码和工作流里的重复执行;它按我给的结构和事实口径产出初稿,我再逐处核对。 AI took on research organisation, copy and visual drafts, page code, and the repetitive execution inside the workflow; it produced first drafts against the structure and facts I gave it, and I checked each one.
我负责的判断What I decided
我负责定义每个任务要解决什么、用什么形式呈现、事实和职责对不对、视觉层级合不合理;海报初版里 AI 曾用错角色名称、把合伙人职责缩窄成「招募」、生成不符合要求的二维码,我对照资料逐页改,并决定最终采用哪一版。 I defined what each task needed to solve, how it should be presented, whether facts and responsibilities were right, and whether the visual hierarchy made sense. In the first poster drafts AI used the wrong role name, narrowed the partner role to “recruitment” and generated a non-compliant QR code; I fixed each page against the source material and decided which version to keep.
四周里做了三件相互接续的事Three connected pieces of work over four weeks
每一件都遵循同一个习惯:先拆任务、定事实口径,再让 AI 产出初稿,我逐处核对,最后把有效做法留下来。Each followed the same habit: break down the task and fix the facts, let AI draft, check every part, then keep what worked.
-
理解团队,参与小组页面Understand the team; contribute to the team page
阅读 34 名成员已开放的名片墙和个人说明书,把共性和特点整理成文字;再和 AI 一起确定小组网站「核心特点」模块怎么呈现,完成初稿。团队最终成果沿用了大部分方向。这是我的初稿,不是整个小组页面的独立完成。Read the shared profiles and personal guides of 34 members and wrote up the shared traits; then worked out with AI how the team site's “core traits” module should present them and produced a first draft, most of whose direction the team kept. This is my draft, not sole authorship of the whole page.
-
洞察转招募物料Turn insight into recruitment materials
检索并整理 9 条小红书定性样本,理解目标学生在意什么;再根据会议纪要设计 8 张招募海报的内容结构,覆盖注意吸引、学生痛点、角色解释、项目流程与成果、人群筛选和报名方式;指挥 AI 生成视觉后逐页核验。因任务调整未实际发布,所以不涉及任何传播数据。Collected and organised nine qualitative Xiaohongshu samples to understand what target students care about; then designed the content structure of eight recruitment posters from the meeting notes — attention, student pain points, the role, the project process and outcomes, audience fit and how to sign up — directed AI to generate the visuals and checked every page. Plans changed and the set was never posted, so there are no distribution metrics.
-
把求职准备搭成工作流Build job-search prep into a workflow
做了一个 KISS 结营复盘模板;把岗位分析、经历匹配、简历事实核验、面试准备和面试后复盘连成一条流程,并加入「面试官视角」检查;反复使用的部分封装成两个可复用的 Skill,存进个人知识库。这条工作流单独有一页说明。Made a KISS debrief template; connected role analysis, experience matching, résumé fact-checking, interview prep and post-interview debrief into one flow, with an interviewer-perspective check; packaged the repeated parts into two reusable Skills stored in a personal knowledge base. That workflow has its own page.