Dibya R. Mishra
Case Studies

Programs that moved the needle

Selected case studies with the architecture choices, execution strategy, and business outcomes that defined them.

AI & Agentic AI

Enterprise Agentic Automation at Scale

A multinational needed to automate 40+ back-office workflows under strict compliance and audit constraints.

Architecture Approach

Cell-based agent runtime with planner/executor agents, deterministic tool layer, policy guardrails, and full observability.

Execution Strategy

Quarterly thin-slice releases; each workflow shipped behind a kill switch with shadow-mode and progressive rollout.

Technology Stack
Agentic AILLM IntegrationAzureKubernetesPythonMLOps
AI / Cloud / Data Components
  • Multi-agent orchestration
  • Hybrid RAG over policy corpora
  • Human-in-the-loop approvals
  • Audit-grade traceability
Outcome & Business Impact

72% cycle-time reduction, $4.1M annual savings, zero compliance incidents across 11 months in production.

Lessons Learned
  • Start with workflows where humans already follow checklists.
  • Invest in evaluation infrastructure before scaling agents.
  • Guardrails are a product surface, not a bolt-on.
Cloud & DevSecOps

SaaS Cloud Modernization for a Global ISV

Legacy monolith couldn't meet enterprise SLAs or regional data residency demands.

Architecture Approach

Strangler-fig decomposition into bounded contexts, cell-based deployment per region, IDP for product teams.

Execution Strategy

18-month program with parallel runs, automated data migration, and weekly stakeholder demos.

Technology Stack
AWSKubernetesTerraform (IaC)GitHub Actions.NET Core / C#Node.js
AI / Cloud / Data Components
  • Internal developer platform
  • Region-aware data plane
  • Progressive delivery
  • FinOps guardrails
Outcome & Business Impact

40x deploy frequency, 38% infra cost reduction, 99.99% availability, 7 new regions in 12 months.

Lessons Learned
  • Modernization is an organizational change as much as a technical one.
  • Pay down platform debt early to unlock product velocity.
  • Treat cells as a product with versioned contracts.
BFSI & Payments

Biller Advantage — BFSI Payments Onboarding at Scale

Merchant onboarding took 3 months across 20+ US states, capping growth for a multi-portal payments platform.

Architecture Approach

Multi-portal platform sharing a hardened payment engine; automated merchant onboarding pipeline with rules-driven KYC and routing.

Execution Strategy

Domain-driven slices per portal; release trains aligned to regulator windows; SRE-led reliability program targeting 99.9% uptime.

Technology Stack
.NET Core / C#SQL / NoSQLREST APIsAzureDevSecOps
AI / Cloud / Data Components
  • Merchant onboarding pipeline
  • Core payment engine
  • Boarding & Servicing platform
  • Reconciliation & audit
Outcome & Business Impact

Merchant onboarding compressed from 3 months to 4 hours (-98%); 99.9% uptime on Boarding & Servicing across the engagement.

Lessons Learned
  • Onboarding velocity is a product KPI — instrument it end-to-end.
  • Idempotency and retries belong in the platform, not the apps.
  • Regulator-aligned release trains beat ad-hoc deployments in BFSI.
Discuss a similar program