Authors: Karthik Reddy

Abstract: Modern enterprise system design has undergone a significant transformation with the widespread adoption of cloud computing, containerization, and automation technologies. Traditional monolithic and on-premise infrastructures are increasingly being replaced by distributed, cloud-native architectures that emphasize scalability, resilience, and rapid delivery. Cloud platforms provide elastic infrastructure and global availability, while containerization technologies such as Docker and orchestration systems like Kubernetes enable portability, efficient resource utilization, and microservices-based design. Automation practices, including DevOps, Continuous Integration/Continuous Deployment (CI/CD), and Infrastructure as Code (IaC), further enhance deployment speed, operational consistency, and system reliability. This review examines the architectural evolution of enterprise systems, key enabling technologies, integration strategies, security considerations, and operational challenges associated with cloud-native environments. Additionally, it highlights emerging trends such as serverless computing, GitOps, and AI-driven infrastructure management. By synthesizing current practices and technological advancements, this paper provides a comprehensive overview of modern enterprise system design and outlines strategic considerations for organizations pursuing digital transformation.

DOI: http://doi.org/10.5281/zenodo.18670037

Published by: vikaspatanker

Authors: Indra Kumar

Abstract: The rapid evolution of cloud computing, distributed systems, and software-defined networking (SDN) has fundamentally reshaped how modern enterprises design, deploy, and manage digital infrastructure. The transition from monolithic, on-premises environments to cloud-native architectures, microservices-based applications, and virtualized networking frameworks has significantly increased scalability, flexibility, and innovation velocity. However, this technological advancement has also introduced substantial operational complexity. Traditional siloed operational models—where cloud operations, application development, and network engineering teams function independently—are increasingly inadequate for managing highly dynamic, distributed, and interconnected ecosystems. In conventional IT environments, infrastructure provisioning, application deployment, and network configuration were handled as separate processes with distinct tools, workflows, and performance metrics. This fragmentation often resulted in delayed deployments, inconsistent policy enforcement, limited cross-layer visibility, and heightened vulnerability to outages or security breaches. As organizations adopt multi-cloud strategies, container orchestration platforms, and API-driven networking, the lack of integration between operational domains becomes a critical bottleneck to agility and resilience.Integrated DevOps practices address these limitations by unifying cloud, application, and network operations under a shared framework of automation, Infrastructure as Code (IaC), Continuous Integration/Continuous Deployment (CI/CD), and end-to-end observability. Rather than focusing solely on collaboration between development and operations teams, integrated DevOps extends its principles across the entire technology stack. This approach promotes shared ownership, policy-driven governance, and cross-functional accountability, enabling organizations to manage infrastructure, code, and connectivity as programmable, version-controlled assets. Automation plays a central role in this integration. Through declarative configurations and orchestration pipelines, provisioning and deployment processes become reproducible and scalable. Advanced observability frameworks—encompassing metrics, distributed tracing, and log aggregation—provide real-time visibility into system behavior across infrastructure, application, and network layers. Additionally, embedded DevSecOps practices ensure that security controls, compliance requirements, and vulnerability assessments are integrated directly into development and operational workflows. This review comprehensively examines the conceptual foundations, enabling technologies, and architectural patterns underpinning integrated DevOps for cloud, application, and network operations. It analyzes key methodologies such as GitOps, which leverages version control systems as the single source of truth; NetDevOps, which introduces automation and programmability into networking; and AIOps, which applies machine learning to operational data for predictive analytics and automated remediation. Furthermore, the review explores implementation challenges, including cultural resistance, toolchain complexity, governance concerns in multi-cloud environments, and skill gaps in cross-domain expertise. By synthesizing current practices and emerging innovations, this study highlights how integrated DevOps serves as a strategic enabler of operational resilience, scalable infrastructure management, and continuous digital transformation.

DOI: http://doi.org/10.5281/zenodo.18670032

Published by: vikaspatanker

Authors: Ganesh K. Bharaskar, Jayesh S. Chavan, Yash K. Pawar, Sagar R. Girase, Prof. Mohan T. Patel

Abstract: The imperative for minimizing response time in vehicular accident scenarios necessitates the development of robust, automated detection and notification systems. Conventional methods often rely on manual intervention, introducing critical delays that severely impact victim outcomes. This paper presents the architecture and performance evaluation of an AI-Based Vehicle Crash Detection and Emergency Notification System (AVC-DENS), designed to provide instantaneous, location-aware alerts upon the occurrence of a significant vehicular impact event. The AVC-DENS employs a tightly integrated Internet of Things (IoT) framework centered around the ESP32 microcontroller unit. Crash detection is predicated upon the real-time analysis of data streams derived from integrated vibration and inertial sensors, complemented by GPS modules for precise spatial localization. Upon algorithmic confirmation of a crash event, the system executes a multi-faceted notification protocol: captured video evidence and temporal-spatial coordinates are immediately transmitted to a secure cloud platform, specifically utilizing Firebase for reliable data persistence and retrieval.

DOI: http://doi.org/10.5281/zenodo.18668675

Published by: vikaspatanker

Authors: Mrs. Yogita P. Gawde, Ms. Richa Mishra, Ms. Akansha Yadav, Ms. Isha Raghwani, Ms. Pritam Anil Yadav

Abstract: Agriculture is the backbone of many developing economies and plays a crucial role in ensuring food security. Traditional farming methods are largely dependent on manual observation and farmer experience, which may lead to inefficient use of water, fertilizers, and other resources. With the emergence of Internet of Things (IoT) and Artificial Intelligence (AI), agriculture is transforming into a data-driven domain. This paper proposes an IoT Simulated Agriculture Platform for Data Driven Farming that integrates simulated sensor data with real-time weather information and AI- based analytics. The platform provides intelligent recommendations related to crop selection, irrigation scheduling, fertilizer management, and soil health improvement. By using simulated IoT data, the system eliminates the need for expensive physical sensors, making it suitable for small-scale farmers, students, and researchers. Experimental evaluation shows that the system effectively supports decision making and promotes sustainable agricultural practices.

Published by: vikaspatanker

Authors: Mr. Khartode V. M., Prof. P. G. Sarasambi, Prof. S. P. Godase, Dr. S. D. Shinde

Abstract: The steering knuckle is a unique component that links the suspension, steering, braking systems, and wheel hub to the vehicle chassis. It bears vertical loads and is crucial for directional control. Given the diverse loads encountered in various situations, it is imperative to ensure high quality, durability, and precision without affecting the steering performance or the vehicle's overall behavior. In the automotive sector, reducing fuel consumption and achieving lightweight designs are critical requirements. A lighter steering knuckle improves performance and reduces production costs. This study aimed to optimize the material used for the steering knuckle joint. Currently, it is constructed from spheroidal cast iron, which provides good strength but is heavy and less resistant to corrosion than other materials. Thus, selecting a material with improved corrosion resistance and lower weight is necessary. The proposed approach investigates the use of Al matrix composites. Initially, the knuckle was designed analytically using mathematical equations. Subsequently, FEA was conducted for all alternative materials, and material optimization was performed. An experimental investigation was conducted to validate the results obtained from the FEA. Keywords: Steering Knuckle, Optimization, FEA, Matrix Composites, Lightweight Design.

DOI: https://doi.org/10.5281/zenodo.18657599

Published by: vikaspatanker

Authors: Krishna Prisad Bajgai, Dr. Bhojraj Ghimire, Niraj Kumar Shah

Abstract: Artificial Intelligence (AI)-driven Financial Technology (FinTech) systems have emerged as transformative tools for enhancing financial inclusion and strengthening risk management in financial institutions. In developing economies such as Nepal, Microfinance Institutions (MFIs) play a critical role in poverty alleviation and access to finance but continue to face challenges related to credit risk, fraud, operational inefficiencies, and limited outreach to underserved populations. This systematic review synthesizes existing empirical and theoretical literature on AI-enabled credit scoring, fraud detection, explainable AI, and regulatory governance frameworks in financial services, with a specific focus on applicability to microfinance contexts. Following PRISMA-based screening and thematic synthesis, 42 peer-reviewed and institutional studies were analyzed. Findings indicate that machine learning models significantly outperform traditional statistical approaches in credit risk prediction and fraud detection, while explainable AI techniques such as SHAP and LIME enhance transparency and regulatory trust. However, substantial gaps remain regarding ethical governance, bias mitigation, and deployment in low-resource microfinance environments. The paper proposes a Nepal-specific conceptual framework aligned with Nepal Rastra Bank (NRB) policies and highlights research directions for responsible AI-driven FinTech adoption in microfinance sectors.

DOI: https://doi.org/10.5281/zenodo.18656780

 

Published by: vikaspatanker

Authors: Snehal Pravin Pangavhane, Mayur Navnath Dhumal, Gayatri virendra patil, Harish Ravindra Badgujar, Mrs. Samiksha Gawali

Abstract: The project entitled “Computer Vision: Realtime object detection using AI and Machine Learning Realtime Eye Strain Detection System”, focuses on enhancing digital well-being by addressing the growing problem of Computer Vision Syndrome (CVS), commonly known as Digital Eye Strain. With the increasing dependency on digital devices for work, study, and entertainment, users often experience symptoms such as eye dryness, irritation, blurred vision, headaches, and reduced concentration. The proposed system utilizes Artificial Intelligence (AI) and Computer Vision (CV) technologies to monitor and analyze real-time indicators of visual fatigue. Using tools such as MediaPipe and OpenCV, it detects parameters like blink rate, eye aspect ratio (EAR), sitting distance, and ambient lighting. A user-friendly PyQt6 graphical interface enables seamless interaction, providing users with real-time alerts, adaptive feedback, and personalized wellness recommendations. By integrating AI APIs like Gemini or Grok, the system generates intelligent insights, preventive suggestions, and health trend reports. This promotes healthy screen habits and reduces the risk of long-term eye strain. The Vision Shield system contributes to digital wellness, productivity improvement, and AI-based health monitoring, offering a scalable solution for students, professionals, and organizations alike.

DOI: https://doi.org/10.5281/zenodo.18656354

Published by: vikaspatanker

Authors: Kasheera Gamith

Abstract: Hybrid fiber–wireless access networks have gained prominence as pragmatic solutions to deployment inefficiencies concentrated in the access segment of broadband infrastructure. Existing research has largely treated hybridization decisions as static design choices made at planning time, based on fixed assumptions regarding cost, performance, and feasibility. However, real-world access networks operate under time-variant conditions, including fluctuating traffic demand, dynamic interference environments, evolving regulatory constraints, and phased infrastructure availability. This paper proposes an adaptive control and dynamic optimization framework for hybrid RF–PON access networks that extends static segment-level substitution models into a time-dependent decision space. By integrating control theory principles, multi-objective optimization, and access network architecture models, the paper demonstrates how hybrid networks can continuously adjust the degree and location of wireless substitution to optimize deployment efficiency and service performance. The proposed framework redefines hybrid access networks as adaptive systems rather than fixed architectures, enabling resilience and efficiency under real-world variability. The paper contributes a novel analytical foundation for intelligent access network control and provides direction for future implementation and empirical validation.

DOI: http://doi.org/10.5281/zenodo.18655088

Published by: vikaspatanker

Authors: Ms. Kamble P. S, Ms. Midge S. M, Ms. Kardile S. R, Ms. More V. A, Ms. Aadhav kalyani

Abstract: The rapid growth of intelligent transportation systems has increased the need for automated vehicle identification solutions that are both cost-effective and easy to deploy. Traditional license plate recognition systems rely on high-resolution CCTV cameras and powerful processing units, which makes them expensive and unsuitable for small-scale or portable applications. To overcome these limitations, this project proposes a compact and economical License Plate Recognition (LPR) system using the ESP32-CAM module. The ESP32-CAM is an IoT-based microcontroller equipped with an OV2640 camera, onboard Wi-Fi, and sufficient computing capability to capture real-time images. In the proposed system, ESP32-CAM continuously monitors the vehicle’s presence, captures an image at the correct moment, and transmits it wirelessly to a server or computer for further processing. The backend system applies image preprocessing techniques—such as grayscale conversion, noise reduction, edge detection, and contour analysis—to isolate the license plate region. Optical Character Recognition (OCR) is then used to extract alphanumeric characters from the detected plate. This approach significantly reduces hardware cost, wiring complexity, and power consumption compared to conventional surveillance-based LPR systems. The designed setup is highly scalable and can be deployed in applications such as automated parking systems, gated community authentication, security checkpoints, toll management, and vehicle tracking solutions. The project demonstrates the potential of integrating embedded camera modules with machine learning-based OCR algorithms to create an accurate, portable, and low-power license plate recognition system. The results confirm that the ESP32-CAM can serve as a reliable foundation for intelligent vehicular monitoring in both academic research and practical field implementations.

Published by: vikaspatanker

Authors: Mr. Sam Paul. T, Venkata Ramana Lingamgunta, Moovendhan S, Ramanakumar R

Abstract: Stock markets are complex, dynamic, and highly volatile systems influenced by macroeconomic indicators, corporate performance, geopolitical events, and investor psychology. Conventional stock forecasting approaches rely heavily on single predictive models, static technical indicators, or human intuition, which are inadequate in capturing non-linear dependencies, regime shifts, and predictive uncertainty inherent in financial time-series data. These limitations increase investment risk and reduce the reliability of automated trading systems, particularly for retail investors in emerging markets such as the National Stock Exchange (NSE) of India. This paper proposes an AI-driven NSE Stock Forecasting and Risk-Aware Trading Decision Support System that integrates classical machine learning, deep learning, market regime detection, and probabilistic uncertainty estimation within a unified multi-model framework. The system employs Linear Regression, Long Short-Term Memory (LSTM), Bidirectional Gated Recurrent Units (GRU), and Temporal Fusion Transformer (TFT) models for multi-horizon forecasting over one to fourteen days. A market regime detection module classifies market conditions into Bull, Bear, or Sideways states and dynamically adjusts model weights in a regime-aware ensemble mechanism, while Monte Carlo Dropout is utilized to generate ninety-five percent confidence intervals to support risk-aware decision-making. A prototype implementation is developed using Python, TensorFlow/Keras, Scikit-learn, Pandas, and Streamlit, operating on historical NSE OHLCV data enriched with thirty-two technical indicators. Experimental results demonstrate that the proposed ensemble framework outperforms single-model baselines in terms of prediction accuracy, variance reduction, and trading signal reliability. The system delivers interpretable forecasts, confidence bands, and automated BUY, SELL, or HOLD recommendations through an interactive dashboard, making it suitable for investors, traders, analysts, and researchers.

Published by: vikaspatanker