Mock Interviews

​

Overview

​

Starting an Interview Session

​

Question Generation

​

Personalized Question Creation

# Source: backend/mock_interview_engine.py:37
generate_interview_questions(resume_context, job_description, skills, experience, question_count=8)

​

What Makes Questions Personalized

• Resume Context: Your top skills, projects, and experience (up to 2,500 chars)
• Skills List: Technologies extracted from your resume (up to 25 skills)
• Experience Level: Years of experience to calibrate difficulty
• Job Description: Target role requirements (up to 2,500 chars)
• Variation Seed: Ensures unique questions across multiple sessions

​

Question Types

1. Technical: Deep-dive into technologies and concepts
   • Example: “Explain your approach to database query optimization in your current stack”
2. Behavioral: STAR method scenarios and soft skills
   • Example: “Tell me about a time you had to resolve a critical production issue under pressure”
3. Project-Based: Deep dives into your actual work
   • Example: “Walk me through the architecture decisions you made in your [specific project]”
4. Situational: Hypothetical problem-solving
   • Example: “How would you handle a disagreement with a senior engineer about technical approach?”
5. System Design: Large-scale architecture challenges
   • Example: “Design a URL shortening service handling 1 billion requests/day”

​

Difficulty Progression

• First 1/3: Easy questions to build confidence
• Middle 1/3: Medium difficulty for core assessment
• Final 1/3: Hard questions to test advanced knowledge

​

Answer Evaluation

​

Scoring Dimensions

# Source: backend/mock_interview_engine.py:174
evaluate_answer(question, user_answer, resume_context, job_description)

​

Grading Scale

• A (85-100): Excellent - Interview-ready answer
• B (70-84): Good - Minor improvements needed
• C (55-69): Average - Significant gaps to address
• D (40-54): Below Average - Major rework required
• F (<40): Poor - Fundamental concepts missing

​

Detailed Feedback

​

1. Strengths

• “You correctly identified the use of mutex locks and explained the difference from semaphores”
• “Strong use of quantified metrics: ‘40% reduction in query time‘“

​

2. Improvements

• “Missing mention of database indexing strategies. You should have discussed B-tree vs Hash indexes”
• “Answer lacked the ‘Result’ component of STAR. Always quantify outcomes”

​

3. Model Answer

• Written as if by an expert candidate
• Includes concrete examples and metrics
• Demonstrates proper structure (STAR for behavioral, requirements→design→trade-offs for system design)

​

4. Keyword Analysis

• Covered Keywords: Concepts you successfully mentioned
• Missing Keywords: Expected terms you should have included
• Helps identify knowledge gaps

​

Session Summary

# Source: backend/mock_interview_engine.py:295
generate_session_summary(questions, answers, evaluations, job_description)

​

Aggregate Metrics

• Average Score: Overall performance across all questions
• Overall Grade: A-F grade based on average score
• Completion Rate: Percentage of questions answered
• Grade Distribution: Count of A, B, C, D, F responses

​

Performance by Question Type

{
  "technical": 72,
  "behavioral": 85,
  "system-design": 45,
  "project-based": 68
}

​

AI-Powered Insights

• References specific question types and scores
• Honest assessment: “Your behavioral questions scored 85% but system design answers averaged only 45%, indicating strong soft skills but gaps in architectural thinking.”

• Not generic advice - tied to YOUR answers
• Example: “Lacks depth in distributed systems concepts like CAP theorem and eventual consistency”

• Specific resources and approaches
• Example: “Study database normalization forms (1NF-BCNF) using SQL practice on LeetCode”

• Customized to YOUR weaknesses
• Example: “When discussing projects, always include quantified business impact (% improvement, $ saved, users affected)”

• Clear yes/no with the biggest blocker
• Example: “Not yet ready - system design skills need 2-3 months focused practice before senior-level interviews”

​

Real-Time Audio Streaming

​

Voice Response Feature

​

Best Practices

​

For Technical Questions

• State the core concept first, then elaborate
• Use specific terminology from the expected keywords
• Include trade-offs and alternative approaches
• Reference real-world experience when possible

​

For Behavioral Questions

• Always use STAR method: Situation, Task, Action, Result
• Include quantified outcomes: percentages, dollar amounts, user counts
• Be specific about YOUR role (not team’s role)
• Keep it concise: 2-3 minutes max

​

For System Design Questions

• Follow structure: Requirements → Design → Trade-offs → Scaling
• Start with clarifying questions
• Draw mental diagrams (or use whiteboard feature)
• Discuss failure handling and monitoring

​

For Project-Based Questions

• Reference actual projects from your resume
• Explain technical decisions and their business impact
• Discuss challenges faced and how you overcame them
• Mention technologies used and why they were chosen

​

Example Workflow

​

Technical Details

​

AI Model

• Uses Mistral Large via API for question generation and evaluation
• Prompts engineered for FAANG-level interview standards
• Fallback questions available if API fails

​

Evaluation Strictness

• 5-word answers score ≤15/100
• No generic feedback - all strengths must quote specific things you said
• Model answers are standalone (200-250 words, all concepts covered)

​

Session Persistence

• All questions, answers, and evaluations stored
• Review past sessions anytime
• Track improvement over time