sunanzhe2004/shushu-internship-resume-optimizer

Shushu Internship Resume Optimizer is a Python command line toolkit, mostly in Chinese, that takes a pile of raw internship materials (code repos, project write-ups, business notes) and turns them into resume bullets and an interview preparation pack. The author's framing is that the usual problem with intern resumes is not lack of material but too much scattered material that does not condense well, and that asking an AI to "just write me a resume" tends to produce vague, exaggerated, or evidence-thin sentences. Rather than generating a resume directly, the tool runs three main stages. The first is achievement_audit, which reads a sources.json describing where to find each piece of material and breaks long project summaries into individual achievement items, each annotated with metrics, evidence, risk points, and a user_check_flags field that flags sentences that sound machine-generated, boast-heavy, or weakly supported. The second is resume_rank, which takes the audited achievements plus a target job description and ranks them, recommending bullet-style phrasings, calling out which numbers still need to be filled in, and noting which sentences are too generic. The third is interview_pack, which produces STAR-format drafts, a project introduction, expected interview question-and-answer pairs, follow-up answers for the risk points it flagged earlier, and a pre-submission checklist. There is an optional fourth piece called doc_knowledge that lets you load business documents and query them in direct, basic RAG, or knowledge_base modes, useful for filling in upstream and downstream business context that the user may not remember in detail. The README is careful about safety. A red banner at the top tells users not to upload internal company documents, real user data, credentials, or anything that cannot be shared publicly. The minimal example input bundled in the repo lets you verify the commands and the output structure before substituting your own materials. Setup needs Python 3.10 or newer. You clone the repo, create a virtual environment, run pip install -e ".[dev]", and then invoke the three modules in order on either macOS, Linux, or Windows PowerShell. Outputs are written as JSON, Markdown, and HTML files in a reports directory. The repository is a re-focused fork of an upstream project by LiuMengxuan04 at github.com/LiuMengxuan04/shushu-internship-tool, with this fork narrowed to the internship-to-resume use case. The README does not state an explicit license.