Hi, I'm Mani

Led the AI transformation of enterprise financial disclosure and reporting systems. Responsible for taking AI features from design through to production, working across the full stack — from Azure and AWS infrastructure to API surfaces to Angular and React frontend integration.

• Designed and built an Advanced RAG pipeline using Azure AI Foundry, Azure OpenAI, and Azure AI Search — combining hybrid retrieval (keyword + vector), semantic re-ranking, intelligent chunking, and query expansion for high-accuracy compliance document processing.
• Integrated AWS Bedrock (Claude and Titan models) as an alternative inference layer, enabling multi-cloud LLM routing and reducing single-vendor lock-in; deployed inference endpoints via AWS Lambda and API Gateway with IAM-governed access control.
• Engineered a multi-agent reporting system using Semantic Kernel, AutoGen, and LangGraph, with specialized agents for data retrieval (SQL/API), LLM reasoning, and structured report generation — reducing manual reporting effort by 60%.
• Built and integrated MCP (Model Context Protocol) enterprise connectors enabling AI agents to interact with SQL databases, REST APIs, and external SaaS systems with natural language queries and dynamic function calling.
• Implemented memory-driven agentic workflows with vector memory and persistent context, enabling multi-turn document-aware AI interactions across long-horizon financial tasks.
• Developed a multi-model routing layer directing Azure OpenAI, AWS Bedrock, and local models to specialized tasks (reasoning, extraction, summarization) to balance cost, latency, and output quality.
• Built scalable .NET 8 microservices integrating Azure OpenAI and AI Foundry APIs, with clean architecture, resilience patterns, and Azure API Management.
• Developed Angular-based frontend modules for AI-assisted report generation and disclosure workflows, consuming AI service APIs with real-time feedback on pipeline progress and output quality.
• Provisioned and managed AWS infrastructure supporting data pipelines: S3 for document storage, RDS (PostgreSQL) for structured data, DynamoDB for agent state and session persistence, and ECS for containerized microservice deployments — all secured via IAM roles and policies.
• Created prompt evaluation frameworks to systematically measure and improve LLM output quality, reducing hallucinations across model upgrades.
• Set up and maintained CI/CD pipelines (Azure DevOps) with automated quality gates including test coverage enforcement and static analysis.
• Instrumented services with OpenTelemetry, distributed tracing, and structured logging, providing operational visibility across distributed AI workflows spanning Azure and AWS.

Delivered enhancements to a regulated healthcare platform, integrating AI capabilities and strengthening API security and reliability across patient-facing and clinician-facing systems.

• Integrated Azure Cognitive Services and Azure OpenAI to enable intelligent document processing and clinical decision support features within the platform.
• Built secure, high-throughput APIs using ASP.NET Core and Azure API Management, with OAuth2/OIDC authentication and input validation aligned to healthcare data compliance requirements.
• Developed real-time communication modules (chat and video) using SignalR, integrated into patient and clinician-facing web applications.
• Established automated regression testing pipelines using Cypress, cutting manual QA cycles and improving release stability.

Developed and maintained microservices components for ADP's enterprise payroll and HCM platform serving global clients across multiple countries.

• Built and owned microservices-based platform APIs using .NET Core and Azure, integrating multiple enterprise data sources and downstream systems.
• Designed event-driven workflows using Azure Service Bus and Azure Functions to automate payroll processing tasks, reducing manual operational overhead.
• Improved test coverage and CI/CD pipeline reliability, contributing to a more consistent delivery cadence across the team.

Built end-to-end ERP and CRM solutions across multiple client engagements in Agile delivery teams.

• Built end-to-end ERP and CRM solutions using ASP.NET MVC, Angular, and SQL Server across multiple client engagements.
• Participated in Agile delivery cycles, authored unit tests, and contributed to code reviews.