couragec/llminternskill

LLMInternSkill is a resume and job-search toolkit aimed at people applying for internships at AI companies, particularly roles involving large language model research and engineering. The primary audience is students and early-career candidates who want help making their experience sound clear and technical without overstating what they actually did. The tool works by reviewing a resume alongside supporting materials such as code files, project notes, papers, and awards. It then produces a set of documents: a polished version of each resume bullet, a verdict on whether the language is safe to use given the evidence available, a tailored version of the resume aimed at a specific job description, a list of questions a technical interviewer would likely ask, and suggested answers. The philosophy is that good resume polish makes real experience clearer rather than inflating it. If a candidate wrote a small demo, the tool will not rewrite that as a production system they deployed at scale. One module called Evidence Guard draws a boundary between what a candidate can honestly claim and what would be called out in an interview. Each bullet gets labeled as safe to write, risky, needs more evidence, or should not be written yet. If a claim lacks supporting evidence, the tool suggests specific open-source projects a candidate could study and reproduce to build that evidence before reapplying. Another module called Interview Drill generates the specific follow-up questions an interviewer would ask for each bullet on the resume, along with examples of weak, acceptable, and strong answers. This is meant to help candidates identify which parts of their resume they cannot defend yet. The toolkit installs as a plugin for AI coding tools like Codex and can also be used by pasting prompts directly. It includes a LaTeX resume template for producing a formatted PDF. The README is in both Chinese and English, and the examples focus on common AI internship tracks including search ranking, retrieval-augmented generation, agent systems, and model training.