Authors: S.Kishore Babu, Yarramreddy Abhinaya, Shaik Riyaz, Velamakuru Jhansi, Payardha Sharon Hephzibah

Abstract: In the contemporary digital landscape, establishing a compelling online presence has become an essential prerequisite for professional recognition and career advancement. Despite the proliferation of web development tools and portfolio platforms, the process of creating, personalizing, and deploying a professional portfolio website remains a technically demanding and time-consuming endeavor for many individuals. Craftly emerges as a transformative solution to this challenge — an AI-powered, full-stack web application that automates the end-to-end process of portfolio generation, customization, and live deployment using modern cloud infrastructure. Craftly integrates Google's Gemini AI API to intelligently parse uploaded resumes in PDF format, extracting structured professional data including personal details, skills, work experience, educational background, and project history. This parsed information is used to automatically pre-fill a portfolio editor, dramatically reducing manual data entry. Users may alternatively input their details manually, providing full flexibility in the content creation process. Once satisfied with their portfolio content, users select from nine professionally designed Handlebars-based HTML templates and deploy their portfolio to Amazon Web Services S3 as a static website — all within a single, unified interface. The deployment pipeline leverages Cloudflare Workers and Cloudflare DNS to provide each user with a unique, publicly accessible subdomain, enabling instant sharing of live portfolio URLs without requiring any domain management knowledge from the user. The backend infrastructure is containerized using Docker and deployed on AWS EC2, with Nginx serving as a reverse proxy for the Express.js API server. User authentication is handled via JSON Web Tokens (JWT), and all portfolio data is persisted in MongoDB. Preliminary evaluation of the system demonstrates significant reductions in the time required to create and publish a professional portfolio, with the end-to-end process from resume upload to live deployment achievable in under five minutes. Craftly represents a meaningful convergence of artificial intelligence, cloud computing, and user- centered design — democratizing professional web presence for students, job seekers, and professionals alike.

Published by: vikaspatanker

Authors: Akshay Mahajan, Sahil Hashmi, Tannish Galhate, Anas Dange

Abstract: People's use of smartphones has increased rapidly in today's world, and as a result, a smartphone can be used effectively for personal security or various other protection purposes. On one hand, we get optimistic hope through a list of facts pertaining to woman empowerment, but on the other hand, we are chastised due to the crimes against women. Problems may come from anywhere and anytime, as women are also growing equally like men so for that purpose they have to travel alone at night where ever they go, they have to travel alone in public transport as well, and for that reason we need to understand and solve this problem of women so they also should not feel any fear regarding their safety. BESAFE aims at delivering a simple yet operational elucidation to this problem. BESAFE aims at developing a simple yet effective solution for empowering womanhood as well as installing a sense.

Published by: vikaspatanker

Authors: Jn Chandra Sekhar, P Nagasri, V Nithinreddy, Sk Yaseen, S Praveen Kumar

Abstract: The rapid growth of digital education has exposed critical limitations in existing e- learning platforms, which predominantly rely on static, pre-built content repositories requiring substantial manual creation and maintenance effort. This paper presents EduVoxus, an AI- powered adaptive e-learning platform that integrates OpenAI's GPT-4o-mini model for dynamic content generation with ten machine learning algorithms implemented entirely from scratch, without reliance on external ML libraries such as scikit-learn, scipy, or numpy. The platform offers four distinct AI-driven learning modes: MCQ quizzes with adaptive difficulty, voice-based practice with speech recognition and AI evaluation, theory question generation and an AI chatbot for instant doubt resolution. The ten from-scratch ML algorithms span multiple domains of educational data mining: Exponential Weighted Moving Average (EWMA) for adaptive difficulty adjustment, SM-2 SuperMemo algorithm for spaced repetition flashcard scheduling, TF-IDF with cosine similarity for content-based recommendations, Ordinary Least Squares linear regression for score trend prediction, K- Means with K-Means++ initialization for learner clustering, user-based collaborative filtering with Pearson correlation, Bayesian Knowledge Tracing (BKT) for mastery estimation, Ebbinghaus forgetting curve modeling for optimal review scheduling, first- order Markov chains for study sequence prediction, and Gaussian Naive Bayes for at-risk learner classification. The platform additionally features comprehensive gamification (points, badges, streaks, leaderboards), role-based access control with user approval workflows, course management with study material uploads, community discussion forums with AI-assisted answers, SM-2 scheduled flashcard decks, bookmarkable Q&A, AI-generated study notes and automatic certificate generation. Built with Flask 3.1.3, SQLAlchemy, Bootstrap 5, Chart.js and Web Speech API, with PostgreSQL support for production deployment on Render.com. Comparative analysis demonstrates that EduVoxus offers capabilities not found in any single existing platform including BYJU'S, Coursera, Udemy and Khan Academy.

Published by: vikaspatanker

Authors: ASR Reddy, S. Trisha, R. Venu Gopal, G. Sathvika, M. Venkatesh

 

Abstract: In this paper a new class of adjustable window function is proposed using a combination of Tangent hyperbolic function and Blackman-Harris 4-term window function. To derive the Tangent Hyperbolic Window function, the authors used the scaled independent variable Tangent hyperbolic functions shifted in opposite directions. The proposed window has the advantage of having 4-shape parameters that have lot of flexibility to vary the shape of the window for the desired spectral characteristics. The performance is compared with Hamming, Hanning, Kaiser and Gaussian windows in terms of the First Null Beam Width, Main Lobe Beam Width, Ripple ratio and Sidelobe roll-off ratio for the same window length with other windows presented for comparison. Simulation results show that Tanh window combined with Blackman-Harris window provides better sidelobe roll off characteristics and other spectral metrics that may be useful for some applications such as filter design and beamforming. Moreover, the paper presents the application of the proposed window in the field of array synthesis, and the comparison is performed with Hamming, Hanning, Kaiser and Gaussian windows. The results show that the array design with Tanh- Blackman-Harrish window provides better results in terms of the spectral metrics such as First Null Beam Width, Main Lobe Beam Width, Ripple ratio and Sidelobe roll-off ratio.

DOI: https://doi.org/10.5281/zenodo.19349175

 

Published by: vikaspatanker

Authors: Md Tanvir Ahamed

Abstract: Software bugs remain one of the most persistent challenges in software development, consuming 50-75% of developer time and costing the global economy over $2 trillion annually. This paper presents a multi-language approach to automated bug detection and fixing using the T5-Small transformer model. We construct a dataset of 2,600 real bug examples from Defects4J, BugSwarm, QuixBugs, GitBugs, and 500 novel multi-error examples. The T5-Small model (60M parameters) is fine-tuned with optimal hyperparameters. Our evaluation framework employs seven metrics with mathematical formulations. Experimental results demonstrate 68.46% Normalized Exact Match, 93.74% F1 Score, and 99.55% ROUGE-1. The model performs effectively on both Python (70.0%) and Java (65.0%). All artifacts are released open-source.

DOI:

 

 

Published by: vikaspatanker

Authors: Rajani Devi K, Gowri Sankar R, Gayathri Reddy R, Harsha Vardhan V, Srikanth T

Abstract: In the modern software development landscape, countless developers—particularly students, beginners, and hobbyists—build innovative web applications but fail to deploy them to the internet due to the complexity of traditional deployment processes. Deploying an application requires extensive knowledge of cloud platforms such as AWS, GCP, or Cloudflare, involving technical hurdles including renting and configuring cloud instances, purchasing domains, setting up web servers, and managing infrastructure. This steep learning curve creates a significant barrier to entry, causing many developers to abandon their fully-functional applications at the development stage without ever making them publicly accessible, thereby limiting innovation visibility and preventing developers from building their portfolios.

 

 

Published by: vikaspatanker

Authors: Tushar Chaudhari

Abstract: The public release of ChatGPT in late 2022 marked a turning point in the adoption and academic investigation of Generative Artificial Intelligence (GenAI). This systematic literature review synthesises 39 peer-reviewed studies to evaluate the applications, pedagogical benefits, and governance challenges of GenAI across global educational contexts. The findings identify four primary application domains: personalised intelligent tutoring, automated content creation, multimodal learning materials, and academic research assistance. Synthesis of the evidence reveals substantial improvements in student performance and affective-motivational states, particularly through adaptive scaffolding and real-time feedback. However, these benefits are countered by significant risks involving academic integrity, "hallucinations," and the potential for cognitive over-reliance. Parallel to these pedagogical concerns, the evidence base remains heavily concentrated in higher education and high-income regions, leaving critical gaps in K-12 settings and the Global South. This review concludes that while GenAI offers transformative potential for personalised learning, its sustainable integration requires robust institutional policy and longitudinal research into long-term cognitive outcomes.

DOI: https://doi.org/10.5281/zenodo.19331843

Published by: vikaspatanker

Authors: Pushkar Chaudhari, Vaibhav Thakre, Tushar Chaudhari, Tanya Bhaute, Dr. Rais Khan

Abstract: Port scanning is a fundamental technique used in cybersecurity for identifying active services and potential vulnerabilities in networked systems. As modern networks grow in complexity, efficient and scalable scanning tools become increasingly important for administrators and security researchers. This paper presents the design and development of an advanced multithreaded port scanner implemented in Python and executed on the Linux operating system. The proposed system aims to provide efficient port discovery, faster scanning performance, and structured reporting for network security analysis. Unlike traditional sequential scanners, the proposed approach utilizes parallel execution techniques to analyze multiple ports simultaneously. The architecture includes modules for user input handling, scanning engine management, multithreading coordination, result processing, and reporting. Experimental evaluation demonstrates improved scanning speed and reliability compared to conventional scanning approaches.

Published by: vikaspatanker

Authors: Mrs. P.V. Javkar, Mr. Damodhar N Bulbule, Mr. Arya D Tapkir, Mr. Kaivalya R Bhadange, Mr. Devraj A Yadav

Abstract: Social media platforms have become a major part of daily communication, marketing, entertainment, and information sharing. Among them, Instagram is one of the most widely used platforms across the world. However, the rapid growth of Instagram has also led to the creation of a large number of fake accounts. These fake accounts are often used for scams, impersonation, phishing, spam promotion, fake giveaways, misinformation, and fraudulent advertisements. Detecting such accounts has become an important research problem in the field of cybersecurity and social media analysis. Traditional fake account detection systems mainly focus on profile-related information such as follower count, following count, number of posts, account age, and user activity. Although these features are useful, they may fail to detect accounts that hide suspicious content inside images. Many fake Instagram accounts include scam messages, promotional offers, fake links, or misleading text inside profile images, stories, and post images. Such hidden text cannot be effectively analyzed using normal text-based techniques alone. This paper proposes a method for detecting fake Instagram accounts using Optical Character Recognition (OCR). OCR is used to extract text from profile pictures, post images, and other visual content associated with an Instagram account. After text extraction, suspicious keywords, spam patterns, links, and unusual promotional phrases are analyzed. These OCR-based features are combined with profile-level features such as follower-following ratio, posting behavior, account age, username structure, and bio information. Based on these features, the account is classified as genuine or fake. The proposed approach improves the detection of fake accounts by analyzing both textual and visual content. This makes the system more effective in identifying hidden spam techniques used by fake profiles. The paper also discusses methodology, algorithm steps, feature extraction, preprocessing, system architecture, results, limitations, and future scope.

DOI: https://doi.org/10.5281/zenodo.19328726

Published by: vikaspatanker

Authors: Miss Payal D. Bhute, Professor Monika Ingole, Professor Vijayata Dalwankar

Abstract: With the growing prevalence of mental health disorders across the globe, the application of Artificial Intelligence (AI) and Machine Learning (ML) has gained significant attention for early detection, prevention, and intervention. This study explores various AI-based models used for mental health self-assessment, including traditional machine learning techniques such as Support Vector Machines (SVM), Logistic Regression, and Random Forest, as well as advanced deep learning approaches. Furthermore, the paper reviews commonly used datasets and highlights the role of Natural Language Processing (NLP) tools in analyzing user-generated data for identifying mental health patterns. Ethical concerns such as data privacy, bias, and transparency are also discussed, along with the feasibility of deploying these solutions through web-based platforms. The objective of this study is to summarize recent advancements and identify existing research gaps, thereby supporting the development of scalable, accessible, and ethically responsible AI-driven mental health systems.

DOI: https://doi.org/10.5281/zenodo.19328659

Published by: vikaspatanker