arXiv subjects

Saima Rafi

Publications and source records attributed to Saima Rafi.

9 recordsLinked to original sources

arXiv2026

Low-Code Paradox in DevOps: Security and Governance Insights from Practitioners

DevOps has become a dominant paradigm in modern software engineering, while low-code development platforms (LCDPs) are increasingly adopted to streamline software development. The integration of these approaches promises efficiency gains but also raises critical concerns regarding security and governance. Despite their growing use, insufficient attention has been given to the implications of these platforms for security and governance in DevOps environments. This study investigates practitioners perspectives on the security and governance implications of LCDPs in DevOps environments. Twelve semi-structured interviews were conducted with IT professionals experienced in low-code and DevOps practices. The data were analyzed using a grounded theory approach to identify emergent themes. Findings reveal that LCDPs help automate tasks; however, they also increase security risks and governance challenges, highlighting the need for robust practices and a security-conscious culture. This study suggests that the intersection of DevOps and LCDPs requires careful governance and proactive security practices. Addressing these issues is essential for organizations to unlock the potential of LCDPs while safeguarding resilience, compliance, and developer needs.

cs.SE↗

arXiv2025

Polypersona: Persona-Grounded LLM for Synthetic Survey Responses

This paper introduces PolyPersona, a generative framework for synthesizing persona-conditioned survey responses across multiple domains. The framework instruction-tunes compact chat models using parameter-efficient LoRA adapters with 4-bit quantization under a resource-adaptive training setup. A dialogue-based data pipeline explicitly preserves persona cues, ensuring consistent behavioral alignment across generated responses. Using this pipeline, we construct a dataset of 3,568 synthetic survey responses spanning ten domains and 433 distinct personas, enabling controlled instruction tuning and systematic multi-domain evaluation. We evaluate the generated responses using a multi-metric evaluation suite that combines standard text generation metrics, including BLEU, ROUGE, and BERTScore, with survey-specific metrics designed to assess structural coherence, stylistic consistency, and sentiment alignment.Experimental results show that compact models such as TinyLlama 1.1B and Phi-2 achieve performance comparable to larger 7B to 8B baselines, with a highest BLEU score of 0.090 and ROUGE-1 of 0.429. These findings demonstrate that persona-conditioned fine-tuning enables small language models to generate reliable and coherent synthetic survey data. The proposed framework provides an efficient and reproducible approach for survey data generation, supporting scalable evaluation while facilitating bias analysis through transparent and open protocols.

cs.CL↗

arXiv2025

Residual GRU+MHSA: A Lightweight Hybrid Recurrent Attention Model for Cardiovascular Disease Detection

Cardiovascular disease (CVD) remains the leading cause of mortality worldwide, underscoring the need for reliable and efficient predictive tools that support early intervention. Traditional diagnostic approaches rely on handcrafted features and clinician expertise, while machine learning methods improve reproducibility but often struggle to generalize across noisy and heterogeneous clinical data. In this work, we propose Residual GRU with Multi-Head Self-Attention, a compact deep learning architecture designed for tabular clinical records. The model integrates residual bidirectional gated recurrent units for sequential modeling of feature columns, a channel reweighting block, and multi-head self-attention pooling with a learnable classification token to capture global context. We evaluate the model on the UCI Heart Disease dataset using 5-fold stratified cross-validation and compare it against classical methods such as Logistic Regression, Random Forest, and Support Vector Machines, as well as modern deep learning baselines including DeepMLP, convolutional networks, recurrent networks, and Transformers. The proposed model achieves an accuracy of 0.861, macro-F1 of 0.860, ROC-AUC of 0.908, and PR-AUC of 0.904, outperforming all baselines. Ablation studies confirm the individual contributions of residual recurrence, channel gating, and attention pooling. t-SNE visualizations further indicate that the learned embeddings exhibit clearer separation between disease and non-disease classes compared to raw features. These results demonstrate that lightweight hybrid recurrent and attention-based architectures provide a strong balance between accuracy and efficiency for clinical risk prediction, supporting deployment in resource-constrained healthcare settings.

cs.LG↗

arXiv2025

Executable Governance for AI: Translating Policies into Rules Using LLMs

AI policy guidance is predominantly written as prose, which practitioners must first convert into executable rules before frameworks can evaluate or enforce them. This manual step is slow, error-prone, difficult to scale, and often delays the use of safeguards in real-world deployments. To address this gap, we present Policy-to-Tests (P2T), a framework that converts natural-language policy documents into normalized, machine-readable rules. The framework comprises a pipeline and a compact domain-specific language (DSL) that encodes hazards, scope, conditions, exceptions, and required evidence, yielding a canonical representation of extracted rules. To test the framework beyond a single policy, we apply it across general frameworks, sector guidance, and enterprise standards, extracting obligation-bearing clauses and converting them into executable rules. These AI-generated rules closely match strong human baselines on span-level and rule-level metrics, with robust inter-annotator agreement on the gold set. To evaluate downstream behavioral and safety impact, we add HIPAA-derived safeguards to a generative agent and compare it with an otherwise identical agent without guardrails. An LLM-based judge, aligned with gold-standard criteria, measures violation rates and robustness to obfuscated and compositional prompts. Detailed results are provided in the appendix. We release the codebase, DSL, prompts, and rule sets as open-source resources to enable reproducible evaluation.

cs.AI↗

arXiv2025

Strategic Motivators for Ethical AI System Development: An Empirical and Holistic Model

Artificial Intelligence (AI) presents transformative opportunities for industries and society, but its responsible development is essential to prevent unintended consequences. Ethically sound AI systems demand strategic planning, strong governance, and an understanding of the key drivers that promote responsible practices. This study aims to identify and prioritize the motivators that drive the ethical development of AI systems. A Multivocal Literature Review (MLR) and a questionnaire-based survey were conducted to capture current practices in ethical AI. We applied Interpretive Structure Modeling (ISM) to explore the relationships between motivator categories, followed by MICMAC analysis to classify them by their driving and dependence power. Fuzzy TOPSIS was used to rank these motivators by importance. Twenty key motivators were identified and grouped into eight categories: Human Resource, Knowledge Integration, Coordination, Project Administration, Standards, Technology Factor, Stakeholders, and Strategy & Matrices. ISM results showed that 'Human Resource' and 'Coordination' heavily influence other factors. MICMAC analysis placed categories like Human Resource (CA1), Coordination (CA3), Stakeholders (CA7), and Strategy & Matrices (CA8) in the independent cluster, indicating high driving but low dependence power. Fuzzy TOPSIS ranked motivators such as promoting team diversity, establishing AI governance bodies, appointing oversight leaders, and ensuring data privacy as most critical. To support ethical AI adoption, organizations should align their strategies with these motivators and integrate them into their policies, governance models, and development frameworks.

cs.SE↗

arXiv2022

Quantum Software Engineering: A New Genre of Computing

Quantum computing (QC) is no longer only a scientific interest but is rapidly becoming an industrially available technology that can potentially tackle the limitations of classical computing. Over the last few years, major technology giants have invested in developing hardware and programming frameworks to develop quantum-specific applications. QC hardware technologies are gaining momentum, however, operationalizing the QC technologies trigger the need for software-intensive methodologies, techniques, processes, tools, roles, and responsibilities for developing industrial-centric quantum software applications. This paper presents the vision of the quantum software engineering (QSE) life cycle consisting of quantum requirements engineering, quantum software design, quantum software implementation, quantum software testing, and quantum software maintenance. This paper particularly calls for joint contributions of software engineering research and industrial community to present real-world solutions to support the entire quantum software development activities. The proposed vision facilitates the researchers and practitioners to propose new processes, reference architectures, novel tools, and practices to leverage quantum computers and develop emerging and next generations of quantum software.

cs.SE↗

arXiv2022

A Vision of DevOps Requirements Change Management Standardization

DevOps (development and operations) aims to shorten the software development process and provide continuous delivery with high software quality. To get the potential gains of DevOps, the software development industry considering global software development (GSD) environment to hire skilled human resources and round-the-clock working hours. However, due to the lack of frequent communication and coordination in GSD, the planning and managing of the requirements change process becomes a challenging task. As in DevOps, requirements are not only shaped by development feedback but also by the operations team. This means requirements affect development, development affects operations and operations affect requirements. However, DevOps in GSD still faces many challenges in terms of requirement management. The purpose of this research project is to develop a DevOps requirement change management and implementation maturity model (DevOps-RCMIMM) that could assist the GSD organizations in modifying and improving their requirement management process in the DevOps process. The development of DevOps-RCMIMM will be based on the existing DevOps and RCM literature, industrial empirical study, and understanding of factors that could impact the implementation of the DevOps requirement change management process in the domain of GSD. This vision study presents the initial results of a systematic literature review that will contribute to the development of maturity levels of the proposed DevOps-RCMIMM.

cs.SE↗

arXiv2022

Classical to Quantum Software Migration Journey Begins: A Conceptual Readiness Model

With recent advances in the development of more powerful quantum computers, the re-search area of quantum software engineering is emerging. Quantum software plays a critical role in exploiting the full potential of quantum computing systems. As a result, it has been drawing increasing attention recently to provide concepts, principles, and guidelines to address the ongoing challenges of quantum software development. The importance of the topic motivated us to voice out a call for action to develop a readiness model that will help an organization assess its capability of migration from classic software engineering to quan-tum software engineering. The proposed model will be based on the existing multivocal literature, industrial empirical study, understanding of the process areas, challenging factors and enablers that could impact the quantum software engineering process. We believe that the proposed model will provide a roadmap for software development organizations to measure their readiness concerning to transformation from classic to quantum software engineering by suggesting best practices and highlighting important process areas, challeng-es, and enablers.

cs.SE↗

arXiv2022

Practitioners Perspective on Motivators of Agile in Global Software Development

In modern software development world, experts are trying to provide the best solutions to their clients. To achieve this, the organizations opt for the agile software development process as it enables them to develop and deliver the product in-time and as per clients expectations. Consequently, in software engineering industry, the Global Software Development (GSD) is the most widely considering software development paradigm as it offers significant strategic and business gains. Seeking the benefits of GSD, the European software engineering organizations are outsourcing their development activities in developing countries. Considering the criticalities of agile adoption in GSD, this work empirically studies the motivators that could positively influence the execution of agile-based GSD in European software industry. A quantitative survey was conducted and data from 139 practitioners working in agile and GSD based projects was collected. The collected observations were further analyzed using Smart-PLS (3.0). The results show that the identified motivators are important to consider by industry experts to successfully apply the agile practices in GSD context.

cs.SE↗