timothyhabashy/interviewos

InterviewOS — AI mock-interview practice for under-resourced students. Built for the Spring Sprint Hackathon, Economic Empowerment & Education track. # InterviewOS **Practice high-stakes interviews when you do not have insider access.** Built for the Spring Sprint Hackathon — *Economic Empowerment & Education* track. --- ## What it does InterviewOS is a structured AI mock-interview *environment* — not a chat window. It plans an interview, asks adaptive questions through an animated interviewer, supports both qualitative *and* technical multiple-choice questions (with rendered LaTeX for math), runs a per-question timer, scores the whole transcript on a 10-dimension rubric, and gives a question-by-question review plus targeted drills the user can take home. It is a single Streamlit app. No database, no auth, no file upload. ## Who it helps Students who have talent but not always access to realistic interview practice: - First-generation college students - Low-income students - Rural and small-town students - Community college students preparing to transfer - Immigrant students - Anyone applying to internships, scholarships, research programs, transfer admissions, first jobs, technical SWE roles, quant roles, scientific computing roles, or data science roles without an alumni network or family Rolodex ## Why it matters Many students do not lose opportunities because they lack ability. They lose them because their **first real interview is also their first serious practice**. Wealthier and better-networked students get to rehearse with parents, family friends, alumni, recruiters, or paid coaches. InterviewOS gives every student a realistic, low-stakes place to practice, get feedback, and try again. ## Features - **Animated interviewer avatar** with floating + blinking animations and a speech bubble — the question feels asked, not pasted. - **Auto, Qualitative, Technical, and Mixed** interview modes. In Mixed mode, qualitative and technical questions alternate. - **Auto mode** plans the interview for the user: it picks the mode, difficulty, question count, and timer based on the opportunity descript …