AskHarder is a free, open-source mock interviewer I built for developers. Paste a job description, take an interview tailored to the role, and get a report with scores, evidence from your answers, and specific points to work on.
I wanted feedback I could check. Each question has an answer key prepared before the interview starts. The report shows which points you covered, what you missed, and what a stronger answer could include.
Practicing with AskHarder
You can add resume highlights alongside the job description and choose a session with three, five, or seven core questions. The interviewer streams its responses and asks follow-up questions when it needs more detail.
After the interview, the report scores correctness, depth, structure, and communication. It includes quotes from your answers, a rubric checklist, and suggested answers. You can track skill scores over time, practice weaker areas, and compare a retake with an earlier attempt.
The scores are practice feedback. They do not predict whether you will pass a real interview.
Building feedback I could inspect
I separated the interviewer from the judge. The interviewer receives questions without answer keys; the judge uses those keys to assess your responses. Python code checks that evidence quotes appear in the transcript and that reported missing points belong to the answer key. It also computes the final pass, borderline, or no verdict from the stored scores.
I built an evaluation suite to check the judge’s scoring order, consistency, evidence quotes, and use of the rubric. The app publishes the results on its methodology page.
The backend uses Python, FastAPI, and Postgres, with React and TypeScript on the frontend. A mock model backend lets developers run the interview flow locally without API keys.
The live app may show an error if my API credits run out. Please try again later.