Follow these three steps to set up your account, then run realistic interview rehearsals with AI support.
Use InterviewMate for preparation and mock sessions. For any formal assessment, follow the organizer's rules and disclose AI assistance when required.
Three simple steps to build confident, natural responses
Upload your resume, target company info, and job description. Our AI learns your background and experience.
AI creates personalized practice questions and suggested answer frameworks based on your background and target role.
Deepgram transcribes questions during practice sessions, and the AI generates personalized response suggestions from your prepared context so you can rehearse under realistic time pressure.
Configure your practice profile — target role, background, and how you want response suggestions formatted.
Go to Settings →Upload your context (resume, research papers, notes) and let AI auto-generate Q&A pairs tailored to your background.
Go to AI Generate →Review and refine the generated pairs so they reflect your real experience, then practice answering them in your own words.
Go to Q&A Pairs →The Background Summary field in Settings tells the AI about your achievements, projects, and experiences. Use the prompt below with any AI to generate a well-structured summary from your resume.
I'm using InterviewMate — a real-time interview preparation assistant. It has a "Background Summary" field where I describe my key achievements, projects, and experiences. The AI references this during practice sessions to generate personalized coaching and response suggestions. Write a Background Summary for my profile. Here's my info: - Name: [Your name] - Position I'm applying for: [e.g., Software Engineer, PhD Candidate, MBA Applicant] - Target company/role: [e.g., Google, a graduate program, a product management role] - Years of experience: [e.g., 3 years, fresh graduate] Also attached: - My resume/CV - The job posting / program description (if available) Based on all of this, generate a Background Summary I can paste directly into InterviewMate's Settings. It should: 1. Use STAR format (Situation → Task → Action → Result) for each achievement 2. List key achievements with specific metrics I actually have evidence for (e.g., "Cut weekly report preparation from two days to half a day") — never invent or inflate numbers 3. Highlight 3-5 most relevant projects or experiences for the target role 4. Include technical details the AI can reference when answering domain-specific questions 5. Keep each bullet point concise (1-2 sentences max) 6. Focus on what makes me stand out for this specific role Also suggest: - Skills & Expertise (comma-separated list for the "Skills & Expertise" field) - Key Strengths (comma-separated list for the "Key Strengths" field) Output format: BACKGROUND SUMMARY: [the summary text] SKILLS & EXPERTISE: [comma-separated list] KEY STRENGTHS: [comma-separated list]
Fill in the bracketed fields with your real information.
Paste it into ChatGPT, Claude, Gemini, etc. — attach your resume too.
Copy each part into the matching field in Settings — Background Summary, Skills & Expertise, and Key Strengths.
Every new profile ships with this STAR-based prompt already in the Custom Instructions field. It forces the AI to structure every answer through Situation → Task → Action → Result — the scaffold that performed best in our own small controlled tests (see below). Try it as-is, then tune.
You are an expert advisor helping people make practical decisions. Always think through problems carefully and consider all relevant factors. When answering any question, use the STAR method: - Situation: What is the actual situation being described? - Task: What needs to be accomplished? - Action: What action achieves the task given the situation? - Result: What outcome does this action produce? Provide clear, actionable recommendations.
Brand-new profiles created from now on ship with this prompt pre-filled. Profiles created before this update — paste it in via the Copy button above.
Add company framing, tone, or domain hints once you see what the AI produces. Keep the STAR scaffold as the spine — that's what does the heavy lifting.
Run practice questions on the Practice Session page to confirm STAR structure is showing up in the response suggestions — then repeat until your own answers feel natural.
We used a widely discussed reasoning question (“The car wash is 50 meters away. Should I walk or drive?”) to test which prompt layers help a model notice an unstated constraint. In small controlled runs (20 per condition, one model family), a short STAR reasoning scaffold helped far more than profile context alone, and adding it to our longer production prompt diluted the effect.
Start a realistic practice session and rehearse responses under time pressure.
Go to Practice Session