Skip to content

Author

Arun Meesala

We have 2 of 8 papers

We haven’t gathered this author’s papers yet. Follow them and we’ll fetch their work.

Not the right person? Other researchers publish under this name.

Review Open access 2025

An Enterprise-Scale Retrieval-Augmented Generative Intelligence Platform for Real-Time Financial Document Synthesis and Regulatory Compliance Automation

Financial services enterprises process vast volumes of unstructured documents-loan agreements, compliance filings, audit reports, and customer correspondence-requiring synthesis, summarization, and regulatory cross-referencing under strict accuracy and traceability constraints. Conventional document automation relies on rule-based extraction or single-pass large language model (LLM) summarization, both of which struggle with provenance tracking, multi-document reasoning, and auditability demanded by financial regulators. This paper presents the Enterprise Generative Document Intelligence Platform (EGDIP), a layered system architecture integrating retrieval-augmented generation (RAG), microservices-based orchestration, and containerized deployment to deliver scalable, traceable document synthesis for regulated financial environments. EGDIP organizes processing into four cooperating layers: a data ingestion layer supporting heterogeneous document formats and streaming updates, a retrieval and indexing layer built on vector-based semantic search, an intelligence layer combining a fine-tuned LLM with deterministic compliance-rule validation, and a consumption layer exposing synthesized outputs through RESTful APIs and interactive dashboards. The system was deployed on a containerized Kubernetes environment and evaluated on a corpus of 180,000 financial documents from loan servicing and regulatory filing workflows. Results show a 58.3% reduction in document review time, a 39.7% improvement in cross-reference accuracy relative to keyword-based retrieval baselines, and sub-second retrieval latency at the evaluated scale. These findings demonstrate that combining retrieval-augmented generation with deterministic validation yields a practical, auditable path toward generative AI adoption in compliance-sensitive enterprise settings.

Arun Meesala · 0 citations
Open access 2024

Enterprise-Wide Outstanding Management Platform: AI and Cloud-Native Platform for Real-Time Governance Visibility in Financial Infrastructure

Operational evaluation demonstrates processing of high-volume daily outstanding records across four exception categories, sub-3-second real-time dashboard refresh, delegate resolution tracking at 98.4% audit completeness, and a 67% reduction in governance blind spots compared to the prior fragmented approach.

Arun Meesala · 0 citations

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.