AI-Powered Job Interview Simulator
Abstract
Have you ever found yourself unable to sleep due to the fear of an upcoming interview, wondering if you are well-prepared? Even if you have thoroughly prepared, the anxiety of standing out among others, concerns about facial expressions, dressing style, a cracking voice, stuttering, using filler words like “umm…,” or even body language aspects such as eye contact can be overwhelming. If you struggle with in-person interview practice due to shyness or lack of opportunities, AI can help!
This AI-Powered Job Interview Simulator is an interactive web application designed to assist graduate students, internship seekers, job seekers who have lost employment, professionals looking for a career change, and others in need of structured interview training in real-world settings. By utilising AI, the simulator conducts AI-driven mock interviews, provides real-time feedback, and evaluates performance using facial recognition, natural language processing, and voice/speech analysis. It assesses users’ responses, body language, and confidence levels to deliver personalized feedback, performance scores, and industry-specific recommendations, helping them refine their interview techniques.
Additionally, this project prioritises essential quality attributes such as reliability, extensibility, and security—ensuring accurate evaluations, future enhancements, and data protection for users across various industries.
Author
Name: Rajanya Paul Choudhury
Student Number: 47324000
Functionality
This application will offer a range of fun and useful features that will transform job preparation by making it interactive and accessible to all users. Here are some descriptions below of full system functionalities that would be delivered when the project is fully built:
- AI-Generated Interview Questions: The platform generates interview questions tailored to the specific industry, job role, and experience level of the user. The difficulty level of the questions will be adjusted based on whether the user is applying for a junior or senior position.
- Soft Skills Assessment (Facial and Voice Analysis): The application analyses the user’s tone, speech clarity, body language, and facial expressions, monitoring eye contact, to provide insights into their communication skills.
- Attire/Clothing and Presentation Recommendations: Based on the job role and industry norms, the platform will suggest appropriate interview attire, helping users feel more confident in their appearance. Since physical presentation plays a role in interviews, this feature ensures users dress appropriately for their desired position.
- Document Suggestions: The system will recommend the necessary documents users should bring to an interview based on job requirements and additional materials that can help them stand out.
- Performance Scoring & Personalised Feedback: After the interview simulation, users will receive a performance score out of 100 along with AI-generated feedback. This will help them identify mistakes and improve for real interviews. Additionally, users can track their progress over time through a learning graph, comparing their performance in past mock interviews.
- Integration with Career Platforms: Users can link their profiles from job application and career platforms they are applying to. Based on this data, the AI will generate relevant interview questions.
- Cloud Deployment & Accessibility: The application is cloud-based, ensuring availability 24/7, accessibility, and cross-platform compatibility across multiple devices and platforms for a seamless user experience.
Scope
For this project, defining the Minimum Viable Product (MVP) is crucial as it outlines the initial version of the product, focusing on the core features users need. The MVP helps determine what needs to be built/implemented first, when to deliver it, and how to allocate resources effectively.
A feasible and realistic MVP for our application will include:
- Basic Web App UI and AI Processing with User Authentication: Implement using React.js/Next.js for the frontend and Flask/Django for the backend.
- Basic AI-Generated Interview Questions for 3-4 Common Industries: Covering key industries such as IT, healthcare, and finance.
- Basic Facial and Voice Recognition using ML/AI Models: Analysing confidence, eye contact, and speech clarity using Whisper (OpenAI’s speech-to-text model) for voice transcription and DeepFace (Keras/TensorFlow-based framework) for face analysis, ensuring accurate real-time feedback.
- Basic Scoring System & Text-Based Feedback: Providing initial evaluation and insights.
- Database: PostgreSQL/MongoDB (store user progress, responses, and feedback).
- Cloud Infrastructure and Deployment: Host the web app on AWS/GCP/Azure to ensure scalability and reliability for future expansion.
For the full product, additional features will be implemented as future enhancements. These include more industry-specific questions, advanced AI-driven analysis, AI coaching sessions, and multi-language support.
Quality Attributes
Quality attributes are critical technical requirements that define the overall quality of a product or system, determining how well it meets user needs and how effective and useful it is. The three key quality attributes chosen for this application are:
1. Extensibility
- The modular AI model architecture will support easy updates and the addition of new interview structures/formats, languages, and industries as the job market evolves.
- The cloud-based design will ensure seamless scalability, allowing new features to be integrated without disrupting core functionalities.
2. Reliability
- The AI should provide consistent and accurate analysis for different types of users and different interview formats/performance. This will require strong AI algorithms and a cloud-based deployment to minimise downtime.
- System reliability will be tested through automated assessments and comparison with expert evaluations.
3. Security
- Data privacy protocols and end-to-end encryption are needed as the system will process sensitive interview responses and voice data. Protection of user data, interview responses, recordings, AI evaluations, and score reports is crucial.
- Role-based access control will prevent unauthorized users from accessing sensitive data.
- Strict data protection policies will be enforced to guarantee user privacy and security.
In scenarios where trade-offs between these quality attributes arise, security will have the highest priority due to the sensitivity of user data and privacy concerns. Reliability will have the next highest priority to ensure consistent and trustworthy user experiences, while extensibility will have relatively lower priority as it primarily impacts future enhancements rather than core immediate functionality.
Evaluation
To ensure an optimal interview preparation experience and verify the system’s performance, extensibility, security, and reliability, the following evaluations will be conducted:
1. Extensibility Testing
- Evaluate the system’s ability to incorporate new interview scenarios without causing failures or disruptions to the existing functionalities; the system should have an ability to adapt.
- Ensure updates and feature additions such as adding new industry types do not negatively impact system performance.
2. Reliability Testing
- Compare AI-generated feedback with expert human evaluations to assess the accuracy and reliability of the AI analysis for speech, facial expressions, and other behavioural factors.
- Automated tests will be implemented to verify AI accuracy and system uptime.
- Test the system’s stability by conducting load tests/stress tests under high user traffic conditions to ensure stable, consistent, and reliable performance.
3. Security Testing
- Identify potential vulnerabilities to enhance security measures by running penetration tests.
- Verify that all sensitive data remains secure when stored or transmitted, with a strong emphasis on data encryption and data privacy compliance (e.g., GDPR, HIPAA).
4. MVP Functionality Testing
- Verify that AI-generated interview questions function correctly across different industries.
- Ensure that facial and voice recognition provide accurate feedback on speech clarity, eye contact, and confidence.
- Validate that the AI-feedback system (scoring and text-based feedback) is accurate and properly stored in the database for consistency.