SBShrinivass Arunachalam Balasubramanian

Shrinivass Arunachalam Balasubramanian

*This is an algorithmically generated author record based on publicly available information.

Shrinivass Arunachalam Balasubramanian has a total of 9 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    ACM TechBrief: Artificial Intelligence’s Effects on Open Source TechBrief

    by Arunachalam Balasubramanian Shrinivass, Simson Garfinkel, Josiah Dykstra, Andy Oram, Nina Shamsi, Jonathan M. Smith

  • Open Access

    Beyond Engagement: A Human-Centered Evaluation Framework for LLM-Powered Recommender Systems

    by Shrinivass Arunachalam Balasubramanian

    Recommender systems are at the heart of how individuals are engaged with information, products, and opportunities across the web. Although some recent progress, including the inclusion of large language models (LLMs) within them, has improved personalization and interaction, nonetheless, most state-of-the-art recommender systems are assessed for their performance through metrics largely centered around engagement, such as clicks, dwell time, or conversions. However, such metrics commonly do not address what matters most from a human experience standpoint, such as trust, agency, transparency, and overall satisfaction. In this work, we develop a human-centered assessment framework for LLM-assisted recommenders that reorients evaluation practices in more long-term engagement-focused assessment contexts towards more human-aligned frameworks. To assess the utility of our framework, we develop a brief qualitative assessment of an LLM-assisted recommendation tool that investigates trade-offs between engagement-focused design choices versus more human centered ones through qualitative assessment. These design choices include factors like ranking strategies, diversity concerns, trust mechanisms, and transparency techniques. We thus explore these concepts in more depth within our qualitative assessment framework. Our analysis draws attention to persistent discrepancies between engagement optimization and human-centered outcomes, and demonstrates how LLM-based recommender systems can be evaluated and designed in more socially aligned ways compared to engagement-centric approaches. We end with a discussion of the implications of our results for recommender system design and how a human-centered assessment is crucial to the next generation of socially aligned recommender systems.

  • Open Access

    Framework-aware query orchestration for Angular micro-frontends: a type-safe approach to GraphQL deduplication and performance optimization

    by Shrinivass A. Balasubramanian

    Graph-based micro-frontend architectures often face redundant data fetching, inconsistent schema enforcement, and delayed load times in modular applications. A framework-aware query orchestration system addresses these challenges through compile-time and runtime optimizations. During compilation, Angular component dependencies are mapped to GraphQL schema fields, enabling type-safe access, preventing over-fetching, and validating schema compliance before execution. At runtime, cryptographic query fingerprinting, route-based prefetching, and lifecycle- aware caching consolidate identical requests, reducing application programming interface (API) calls from 8.2 to 3.1 per session (−62%) and improving cache hit rates from 12% to 89%. Evaluations on synthetic benchmarks and a production-grade healthcare dashboard demonstrate reductions in redundant queries (−67%), bundle size (−24%), and Time to Interactive from 3.5 to 1.8 s (−49%), while eliminating Total Blocking Time. Type-safe integration further minimizes runtime errors (−92%) and accelerates schema migrations (−85%), while centralized query management reduces boilerplate code by 67%, improving developer productivity. Resource audits indicate potential payload savings of over 5 MB through compression, minification, and removal of unused assets. In a realistic Angular micro-frontend case study, the orchestrator reduced redundant GraphQL requests and manual query coordination by up to 87% relative to a production-calibrated Apollo Client baseline, with statistically significant improvements in Time-to-Interactive (TTI) under diverse network conditions.

  • Open Access

    ng-reactive-lint: Smarter Linting for Angular Apps

    by Shrinivass Arunachalam Balasubramanian

    Reactivity is central to Angular applications, yet subtle misuse of Observables, Signals, and change-detection often leads to performance regressions that are difficult to diagnose. Although Angular 17 introduced a unified, signal-first model, most enterprise codebases still rely heavily on legacy RxJS patterns that create unpredictable update flows, memory leaks, and excessive change cycles. To address these issues, we developed ng-reactive-lint, a deterministic static analysis tool that understands Angular's component semantics, lifecycle hooks, template bindings, and reactivity patterns. Unlike generic ESLint or RxJS plugins, ng-reactive-lint performs framework-aware analysis to detect high-impact anti-patterns and provide actionable, context-specific fixes. Evaluation across five large real-world projects showed reductions of up to threefold in unnecessary change detection cycles and up to 75% lower peak memory usage. The tool offers a practical, automated path to adopting modern Angular reactivity at scale.

  • Open Access

    Enhancing User Experience in Single Page Applications with Angular: Advanced Techniques in Lazy Loading, State Management, and Component Optimization

    by Shrinivass Arunachalam Balasubramanian

    Modern Single Page Applications (SPAs) face critical performance challenges characterized by three-phase degradation patterns: (1) prolonged initial load times from monolithic bundles, (2) state management complexity in large-scale applications, and (3) UI thread contention from redundant component re-rendering. This paper presents AngularOpt, a novel optimization framework that combines three technical innovations to address these challenges: route-based code splitting with dynamic import thresholds at θ = 85 % viewport proximity, deterministic state transitions through NgRx finite state machines, and component-level optimization via On Push change detection with Zone.js throttling. Empirical evaluation across 10 000 simulated patient records demonstrates statistically significant improvements (p < 0.001), including a 53 % reduction in First Contentful Paint (FCP) from 4.5 s to 2.1 s (σ = ±0.3 s) and 33 % memory footprint reduction from 180 MB to 120 MB (CI95% [117 MB, 123 MB]). Validated through 1200 Lighthouse audits across healthcare, finance, and e-commerce domains, AngularOpt reduces Time to Interactive (TTI) variance by 40 %, translating to 22 % higher user retention and 17 % faster task completion rates in production environments. These improvements address critical needs in time-sensitive applications, such as telemedicine systems that require sub-second EHR access and fintech platforms that demand real-time portfolio updates. The architecture demonstrates linear scalability up to 50,000 DOM elements while maintaining 60 FPS rendering performance, with a novel Ahead-of-Time (AOT) compilation profile balancing build-time constraints (µ = 8.2 min) against runtime gains (∆ = 2.4 s). Currently adopted by three Fortune 500 companies for petabyte-scale data workflows, this work establishes enterprise-ready design patterns for Angular optimization. Future extensions focus on integrating server-side rendering through Angular Universal and implementing machine learning-driven predictive loading for optimizing the dynamic user journey. The code and resources used in this study are available in their GitHub repository: https://github. com/Shrinivas sab/Enhance-SPA.

  • Governai: Policy-Driven Model Governance for Dynamic and Multi-Tenant AI Systems

    by Sana Zia Hassan, Karan Alang, Nagarjuna Nellutla, Shrinivass Arunachalam Balasubramanian, Vamshidhar Morusu

    Platforms, smart cities, and federated ecosystems, where the governance of models needs to be responsible, and secure, that is, policies-compliant. GovernAI suggests an extended structure that allows intelligent adjustment to the policy-based governance of AI models in the life cycle. GovernAI deals with the main issues of fairness, accountability and compliance management among different stakeholders, by adding declarative policy specifications, tenant-aware access controls, model lineage tracking and real-time behavioral audits. This architecture is capable of feature dynamic policy enforcement, automatic model versioning and dispute resolution in shared landscapes. Experimental analyses prove that GovernAI works in real-world situations minimizing the number of policy violation by 45 %, improves auditability in the multi-tenant setting and allows timely action to counter model misbehavior or adversarial attacks. Touching on ethical AI principles, this work is an application of the domain-specific governance layer that makes ethics scalable and interpretable across the whole AI domain and down to the most complex AI infrastructures.


  • Open Access

    Adaptive Hybrid Sort: Dynamic Strategy Selection for Optimal Sorting Across Diverse Data Distributions

    by Shrinivass Arunachalam Balasubramanian

    Sorting is an essential operation in computer science with direct consequences on the performance of large scale data systems, real-time systems, and embedded computation. However, no sorting algorithm is optimal under all distributions of data. The new adaptive hybrid sorting paradigm proposed in this paper is the paradigm that automatically selects the most effective sorting algorithm Counting Sort, Radix Sort, or QuickSort based on real-time monitoring of patterns in input data. The architecture begins by having a feature extraction module to compute significant parameters such as data volume, value range and entropy. These parameters are sent to a decision engine involving Finite State Machine and XGBoost classifier to aid smart and effective in choosing the optimal sorting strategy. It implements Counting Sort on small key ranges, Radix Sort on large range structured input with low-entropy keys and QuickSort on general purpose sorting. The experimental findings of both synthetic and real life dataset confirm that the proposed solution is actually inclined to excel significantly by comparison in execution time, flexibility and the efficiency of conventional static sorting algorithms. The proposed framework provides a scalable, high perhaps and applicable to a wide range of data processing operations like big data analytics, edge computing, and systems with hardware limitations.