MOCK INTERVIEW SYSTEM USING AI

Seminar Report
submitted in
Partial fulfillment of requirement for the award of
Degree of

Bachelor of Technology
in
Computer Science and Design

Submitted by

Adil Dhote - 220707968

Atharva Pathak -

220707890 Atharva Kumare

- 22070716 Devanshu

Bhure - 22070116 Prit

Nimbalkar – 2207006

Guided by
Asst Prof. Sumedha R. Chokhandre
Asst.Prof.
Department of Computer Science and Design

Department of Computer Science and Design

Yeshwantrao Chavan College of Engineering

(An autonomous institution affiliated to Rashtrasant Tukadoji Maharaj Nagpur University)

Nagpur - 441 110

November 2024
Prof. Prerna C.
Coordinator, S eminar
Department of Computer Science and Design,
Y.C.C.E, Nagpur
 `Yeshwantrao Chavan College of Engineering,Nagpur

CERTIFICATE

This is Certified that the Seminar report entitled “Mock Interview using AI” has been
successfully completed by Adil Dhote , Devanshu Bhure , Prit Nimbalkar , Atharv Pathak ,
Atharva Kumare under the guidance of Asst Prof. Ms. Sumedha R. Chokhandre in
recognition to the partial fulfillment for the award of the degree of Bachelors of
Technology in Computer Science and Design, Yeshwantrao Chavan College of
Engineering (An Autonomous Institution Affiliated to Rashtrasant Tukdoji Maharaj Nagpur
University).

Guide
Asst.Prof. Sumedha R. Chokhandre
Assistant Professor
Department of Computer Science and Design, Y.C.C.E, Nagpur

Prof. Prerna C. Jawdand Dr. R. C. Dharmik

Coordinator, Seminar Head,
Department of Computer Science and Design, Department of Computer Science and
Y.C.C.E, Nagpur Design , Y.C.C.E, Nagpur
 ACKNOWLEDGEMENT
I wish to express my sincere and deepest gratitude to my guide Ms. Sumedha R.
Chokhandre for her valuable and unique guidance. I would also like to thank her for the
constant source of help, inspiration and encouragement in the successful completion of my
technical seminar. It has been my privilege and pleasure to work under his expert guidance.

I would like to thank Dr. R.C. Dharmik(HOD) for including a technical seminar in
our curriculum which helps me to improve my technical knowledge. I would again like to
thank Dr. U. P. Waghe, Principal (YCCE) who has provided all institutional facilities as
and when I express my thanks to all the staff members of Information Technology
department who have directly or indirectly extended their kind co-operation in the
completion of my technical seminar report.

Name of Students
Adil Dhote(37)

Atharv Pathak(39)

Atharva kumare(40)

Devanshu Bhure(43)

Prit Nimbalkar(58)
 INDEX

Sr.No Content Page No

I Introduction 4
II Literature Review 5
III Analysis for Project scope 7
IV Work Done 8
V Results and Discussions 16
VI Summary and Conclusions 18
VII References 19
 I Introduction
In today's competitive world, job seekers and students looking to improve
their interview skills face a major challenge: the lack of real-time feedback
and practical experience in interview scenarios. While traditional mock
interviews are effective, they are often limited by human availability and
cannot scale to provide personalized practice sessions at any time.

This project presents a Mock Interview System Using AI, an automated

solution that mimics real-world interview experiences. Using AI technologies,
the system allows users to practice interviews by interacting with a virtual
interviewer, receiving instant feedback on their answers, and improving their
responses with the help of tailored suggestions and analysis.

The primary aim of this system is to provide candidates with an interactive

platform to prepare for job interviews, helping them refine their
communication, confidence, and overall interview performance.

With flexible scheduling, you can practice at your convenience, accessing a

comprehensive library of common interview questions and expert tips. Developed in
collaboration with HR professionals and industry experts, [Your System Name] is
designed to build your confidence and polish your skills, ensuring you make a lasting
impression in your next interview. Join the ranks of successful candidates who have
enhanced their interview performance with our AI-powered system. Start your journey
to interview excellence today!
 II Literature Review

A literature review for the topic "Mock Interview System Using AI" involves
examining existing research, tools, technologies, and methodologies related to
AI-driven systems designed for conducting and evaluating mock interviews.
Here's an overview of key points to include in such a review

Introduction to AI in Mock Interviews

AI-powered mock interview systems aim to simulate real-world interview
scenarios, provide feedback, and enhance candidates' skills.These systems
often leverage technologies like Natural Language Processing (NLP), speech
recognition, sentiment analysis, and machine learning models to evaluate
responses and provide insights.AI systems are particularly effective in
automating the preparation process and personalizing learning based on user
performance.

Existing Research and Tools

Studies highlight the use of AI in professional development, particularly in
skill-building applications such as mock interviews and soft skill
enhancement.Platforms like HireVue, Big Interview, and MyInterview use AI
for candidate assessment in real interviews and mock practice.These tools
utilize AI to analyze body language, speech clarity, tone, and content
relevance.Research has shown that AI-driven systems can mimic behavioral
and technical interview questions effectively.

Key Components of AI-Powered Mock Interview

Systems
AI-powered mock interview systems consist of several key components that
work together to create an interactive and insightful experience. Speech
recognition technology plays a vital role by converting spoken responses into
text for further analysis, often leveraging tools like Google Speech-to-Text or
Microsoft Azure Speech Services. Natural Language Processing (NLP) is
employed to analyze the textual content of responses, focusing on aspects
such as grammatical accuracy, fluency, and relevance. This is achieved using
advanced libraries and models like spaCy or BERT, which help evaluate the
quality and substance of answers. Additionally, sentiment and emotion
analysis is integrated to assess confidence levels, stress, or hesitation through
both tone of voice and facial expressions. This component combines audio
sentiment analysis with video-based emotion detection to provide deeper
insights into a candidate’s demeanor. Finally, scoring mechanisms are
implemented to assign scores based on industry benchmarks, clarity, and
technical knowledge, ensuring that feedback is both objective and relevant.
Together, these components enable a comprehensive evaluation of interview
performance.
AI-powered mock interview systems offer several advantages that enhance the
user experience. Personalization is a key feature, as the system adapts the
difficulty level and question types based on the candidate's progress and skill
level, ensuring a tailored learning experience. Immediate feedback is another
critical component, providing actionable insights on areas such as speech
clarity, confidence, and the quality of content, helping users improve in real
time. The systems also promote accessibility, allowing candidates to practice
interviews anytime and anywhere, reducing dependency on human
interviewers. Additionally, they are highly cost-effective, minimizing the need
to hire professional interview coaches or invest in elaborate mock interview
setups, making them a practical solution for users seeking affordable and
efficient preparation tools..

Challenges in AI-Powered Mock Interview Systems

AI-powered mock interview systems face several challenges. Bias in AI
models can lead to unfair evaluations if the algorithms reflect biases present in
the training data. Data privacy concerns arise as user responses, recordings,
and personal information must be securely handled. Additionally, the lack of
realism in AI interactions may fail to capture the subtleties of human
communication. Technical constraints, such as errors in speech recognition or
NLP processing, can result in inaccurate assessments, undermining the
system’s reliability. Addressing these issues is crucial to ensuring fairness,
accuracy, and user trust in such systems.
Future Directions and Innovations
Future advancements in AI-powered mock interview systems include
integration of multimodal analysis, combining voice, text, and video inputs for
comprehensive evaluations. Adaptive learning systems will evolve by
incorporating user feedback and external data, enhancing their effectiveness.
Gamification can make the interview preparation process more engaging and
less intimidating. Systems designed with cross-cultural adaptations will
address communication differences and expectations across regions.
Additionally, advanced AI models, like generative AI (e.g., ChatGPT), will
simulate dynamic and unpredictable interview scenarios, offering a more
realistic and challenging experience. These innovations aim to make AI-
driven mock interviews more robust and user-friendly
 III Analysis of project scope
Mock interview systems powered by Artificial Intelligence (AI) are designed
to prepare individuals for real-world interviews by simulating interview
scenarios, evaluating performance, and providing constructive feedback.
These systems are instrumental in enhancing candidates' confidence,
addressing weaknesses, and improving communication skills.

Preparing for interviews often involves challenges such as limited access to

human mentors or interviewers, subjective feedback influenced by individual
perspectives, and the absence of real-time, personalized, and scalable
evaluation mechanisms. To address these issues, this project leverages AI to
create an intelligent, scalable, and objective mock interview system. The
primary objective is to develop an AI-driven platform that simulates realistic
interview scenarios, provides immediate and actionable feedback on areas
such as communication skills, technical knowledge, and body language, and
enhances accessibility and cost-effectiveness in interview preparation.
Key features of the system include tailored question generation based on user-
selected fields, such as technical or HR interviews, and real-time feedback on
speech clarity, tone, grammar, and body language. Sentiment analysis detects
confidence levels, stress, and emotional states, while performance scoring
generates a detailed scorecard highlighting strengths and improvement areas.
Additionally, the system offers replay functionality, allowing users to review
recorded sessions and self-analyze their performance. Core system
components include a speech recognition module to transcribe responses, an
NLP module for evaluating relevance and coherence, a sentiment and emotion
analysis system using video tools, and a scoring mechanism that synthesizes
data to provide comprehensive feedback via an intuitive user interface.
The workflow involves user registration, interview simulation with tailored
questions, real-time data processing, and feedback generation, culminating in
progress tracking. Despite its strengths—scalability, objectivity, and cost-
effectiveness—the system faces challenges like limited replication of nuanced
human interactions, speech recognition errors, and data privacy concerns.
Opportunities for growth include integration with job portals, multilingual
support, and advanced personalization. Use cases span students preparing for
placements, job seekers refining skills, and organizations training employees.
Expected outcomes include enhanced confidence, improved communication,
and a personalized learning experience, although bias in AI models, privacy
issues, and realism limitations remain challenges to overcome.
 IV Work Done
The Mock Interview System comprises several components that work
together to create a seamless experience:

3.1 User Interface (UI)

 Web-based Interface: A user-friendly interface that allows users to
select interview scenarios, start mock interviews, and view feedback.
 Video and Audio Interaction: The system utilizes webcam and
microphone access for real-time interaction, allowing users to respond
to questions with their voice and facial expressions.
 Results Dashboard: Displays detailed feedback on performance,
including analysis on tone, language, content structure, and body
language.

3.2 Speech-to-Text and NLP Engine

 Speech Recognition: Converts spoken language into text using speech
recognition technologies (e.g., Google Speech-to-Text API, Azure
Speech API).
 Natural Language Processing (NLP): The NLP engine processes user
responses to extract meaning, analyze the structure of responses, and
detect keywords, sentiment, and clarity.
 Sentiment Analysis: Detects emotions in the user's speech (e.g.,
confidence, stress, hesitation) and provides insights into their tone and
mood during the interview.

3.3 AI Interviewer
 AI-Based Question Generator: Uses machine learning algorithms to
generate dynamic interview questions based on job roles, industries, and
the user’s previous answers.
 Dynamic Conversation Flow: Adjusts follow-up questions based on
the user’s responses to maintain a natural interview conversation.
 Personalized Questioning: Based on the user’s performance (e.g.,
knowledge of a subject, confidence in their answers), the AI interviewer
tailors questions to either challenge the candidate or help them feel
more comfortable.
 3.4 Feedback and evaluation system
 Answer Evaluation: After each response, the system evaluates the
answer's relevance, coherence, clarity, and confidence.
 Performance Analytics: Provides feedback on several parameters such
as:

o Content Analysis: Was the answer relevant and well-structured?

o Communication Skills: Were the responses clear and articulate?
o Behavioral Traits: Confidence, tone, eye contact, body language
(if video is available).
o Vocabulary and Grammar: Use of appropriate vocabulary and
grammatical accuracy.
 Overall Feedback Report: Generates a comprehensive report that
highlights strengths and areas for improvement, along with suggestions
for enhancing performance.

3.5 Technology Stack

The system leverages a variety of technologies to build and implement its
functionalities:

 Frontend:
o HTML, CSS, JavaScript (React.js for responsive and dynamic
interfaces)
o WebRTC for real-time video and audio streaming
 Backend:

o Python (Flask/Django) for API development and server-side

logic
o Node.js for handling asynchronous tasks and client-server
communication
o MySQL/PostgreSQL for database management (for storing user
profiles, interview history, feedback, etc.)
 AI and NLP:

o NLP: Spacy, NLTK, or OpenAI GPT (for natural language

understanding)
o Speech-to-Text: Google Cloud Speech-to-Text or Microsoft
Azure Speech Service
o Sentiment Analysis: TextBlob, VADER, or custom-trained
sentiment analysis models
o r predictive modeling and behavior analysis
  Cloud Hosting:

o AWS or Google Cloud for hosting, storage, and AI model

deployment
 Security:

o HTTPS for secure data transmission

o JWT (JSON Web Tokens) for user authentication and
authorization

3.4 Key Features

 Real-Time Feedback: Candidates receive immediate feedback after

each question and response, ensuring continuous improvement.
 Multiple Interview Scenarios: Users can choose from a variety of
mock interview types, including technical, behavioral, and situational
interviews.
 Dynamic Interview Flow: The AI interviewer adjusts the interview
flow in real-time, based on the user’s responses.
 Personalized Tips: Based on the candidate’s strengths and weaknesses,
the system provides specific suggestions for improvement.
 Speech and Sentiment Analysis: Analyzes the tone, pitch, and clarity
of responses, as well as non-verbal cues (if video is enabled).
 Report Generation: After the interview, users receive a detailed report
with metrics on their performance, including language fluency,
response relevance, and overall communication skills.

3.5 Processing:

 TF-IDF Transformation: Converts both answers into numerical

vectors.
 Cosine Similarity: Computes similarity as a number between 0 and 1,
where 1 is an exact match.
 Scaling: Multiplies the similarity by 100 to display it as a percentage.
 Output: Similarity score (as a percentage).
Output :

Question: Why do you want to work here?

Ideal Answer: "I admire the company's commitment to innovation."
User Answer: "I am drawn to your company's focus on

innovation and creativity." Similarity

Score: 91.25%

The script outputs the similarity score after each response, providing a
quantitative measure of alignment with the expected answers.
 V Results and Discussions

Key Highlights

Objective Evaluation:
The system uses cosine similarity to objectively measure the semantic similarity
between the user's answer and the expected answer. This avoids subjective bias in
evaluation.

Feedback Loop:
Users receive instant feedback in the form of a percentage, enabling them to gauge
their performance immediately.

Simplicity and Adaptability:

The code is modular and can easily be extended by adding more questions or refining
the predefined answers.

Practical Application:
This can be used to help candidates prepare for real-world interviews by simulating a
structured feedback mechanism.

Limitations

Hardcoded Questions and Answers:

The questions and their answers are predefined, limiting the flexibility to adapt
dynamically to user inputs or industry-specific scenarios.

Surface-Level Understanding:
Cosine similarity focuses on the semantic match of words, but it cannot understand
deeper context, tone, or creativity in responses.

Input Dependency:
Minor variations in phrasing can lead to significant changes in similarity scores, even
if the core meaning is the same.

Limited Feedback:
While similarity scores indicate alignment, they do not provide actionable feedback
on how to improve.
Potential Improvements

Dynamic Question Bank:

Implement a larger and customizable database of questions to provide a
more comprehensive mock interview experience.

Contextual Understanding:
Enhance the system with natural language processing (NLP) models
(e.g., GPT or BERT) to understand responses better. Result and
Discussion

Results

The script is an interactive AI-driven mock interview system designed

to evaluate user responses against predefined ideal answers using cosine
similarity based on TF-IDF (Term Frequency-Inverse Document
Frequency). Below are key observations and results:

Sample Execution Scenario

Assume the following user inputs during execution:

Question: Tell me about yourself.

Ideal Answer: "I am a software engineer with experience in developing
web applications."
User Answer: "I am a software developer focusing on web application
design and development."
Similarity Score: 89.12%

Question: What are your strengths?

Ideal Answer: "I am adaptable and work well under pressure."
User Answer: "I adapt quickly to new environments and handle stress effectively."
Similarity Score: 85.40%

Detailed Feedback:
Provide suggestions or hints to improve user responses beyond just a similarity
score.

Support for Multi-Language:

Adapt the system for multiple languages to cater to a broader audience.
 VI Summary and Conclusion
The Mock Interview System Using AI represents a significant step forward
in personalized job interview preparation. By integrating speech recognition,
natural language processing, and real-time feedback mechanisms, the system
offers an interactive, scalable, and cost-effective solution for job seekers.

The system not only helps users improve their communication skills but also
provides them with in-depth insights into their interview performance. With
continuous enhancements, such as more advanced AI models and additional
interview types, this system has the potential to revolutionize the way people
prepare for job interviews.

The AI-based mock interview system effectively evaluates user answers in

terms of similarity to predefined responses. It provides a foundational
structure for building more advanced systems capable of evaluating qualitative
aspects of interview answers, offering actionable feedback, and improving
candidates' readiness for actual interviews.

The "Mock Interview System Using AI" is a step toward democratizing

interview preparation. By utilizing advanced AI technologies, this project has
the potential to provide personalized, scalable, and effective solutions for
users, bridging the gap between candidates' skills and employer expectations.
 Future Enhancements
Multilingual Support: Adding support for multiple languages to cater to a
global audience.

Video Analysis: Implementing more advanced facial recognition and body

language analysis to provide feedback on non-verbal cues.

Integration with Job Portals: Allowing users to import job descriptions and
tailor mock interviews specifically for particular roles.

Advanced Personalization: Using machine learning to better understand the

user’s weaknesses and adjust interview difficulty and feedback accordingly.

Personalized Learning Paths

Skill Gap Identification: After each mock interview, the AI can analyze the
candidate's performance, identify weaknesses or gaps in knowledge, and suggest
personalized learning paths or materials (e.g., courses, articles, videos) to
improve.
Tracking Progress Over Time: AI can create a learning and progress
dashboard that tracks improvement, provides data-driven insights, and visualizes
trends in key areas (e.g., communication, technical skills, confidence).
 VII References
1. Speech Recognition API Documentation
2. NLP and Sentiment Analysis Resources
3. WebRTC Overview
4. Here are some references and resources related to the development and
concepts used in the provided mock interview system code:
5. Technical References
6. TF-IDF Vectorization
7. Scikit-learn Documentation:
8. Learn more about TF-IDF and its implementation in Python using Scikit-
learn:
9. https://scikit-
learn.org/stable/modules/generated/sklearn.feature_extraction.text.Tfidf
Vectorizer.html
10.Cosine Similarity
11.Cosine Similarity for Text Comparison:
12.A detailed explanation of how cosine similarity works and its use in
comparing text vectors:
https://www.researchgate.net/publication/361451105_Cosine_Similarit y_-
13._A_Computing_Approach_to_Match_Similarity_Between_Higher_Edu
cation_Programs_and_Job_Market_Demands_Based_on_Maximum_N
umber_of_Common_Words
14.Stack Overflow Discussions
15.On using cosine similarity for text matching:
https://stackoverflow.com/questions/8897593/how-to-compute-the-
similarity-between-two-text-documents
16.General AI Concepts
17.Natural Language Processing (NLP)
18.NLP with Python:
19.A book and online guide covering the basics of text processing and
20.machine learning in Python:
https://www.nltk.org/book/
21.AI in Education
22.Using AI for Training and Mock Interviews:
23.Articles and case studies on how AI is used in skill development and
interview preparation:
24.https://www.edtechmagazine.com/
25.Similar Projects and Tools
26.AI Mock Interview Platforms
27.Pramp:
28.A real-world example of an AI-powered mock interview platform:
https://www.pramp.com/
29.Hiration:
30.How AI mock interviews are used for resume building and interview
preparation:
31.https://hiration.com/
32.Scikit-learn Example Codes
33.Text Feature Extraction with TF-IDF:
34.Official examples and tutorials on text preprocessing and feature
extraction:
35.https://scikit-
learn.org/stable/auto_examples/feature_extraction/plot_document_clust
ering.html
36.If this code or project will be submitted for a formal review, be sure to
37.cite these references accurately following your preferred citation style (e.g.,
APA, IEEE).