Skip to introduction

Building

Engineering manager, tech lead, and senior architect for high-volume fraud and risk decisioning on .NET, Kafka, and AWS. Led a 10,000+ rule migration with zero production defects, took Sev-1 incidents from about 70 a month to zero, and applies LLM and RAG to rule conversion and decision analysis. Still hands-on: writes the code, owns the design, and stays with a production issue until it has a permanent fix. Open to a Tech Lead or Senior Architect role. AWS certified. Based in Pune, India.

  • Design for scaleMulti-region, auto-failover, Kafka-streamed decisioning
  • Modernize safelyLegacy to cloud with zero customer-facing downtime
  • Apply AI where it is provableLLM rule conversion, checked in simulation
  • Lead the teamDesign, TDD, code review, delivery for a team of 10
Talk about a roleGitHubLinkedInResume
  • C# / .NET
  • Architecture
  • Microservices
  • API design
Memoji of Prince Shankushal
  • Sparkling Logic rules migration
  • Risk and fraud platform
  • Vendor gateway
  • Kafka consumers and async processing

Selected impact

The outcomes, and the decision behind each one.

10,000+

Fraud and risk rules migrated from FICO Blaze Advisor to Sparkling Logic in 3 months, zero production defects

HowDependency injection kept the C# orchestrator engine-agnostic, and simulation proved parity before traffic moved.

  • Migration strategy
  • .NET
  • LLM

70 → 0

Sev-1 incidents per month

HowCQRS split reads from writes, and the team fixed causes instead of symptoms.

  • CQRS
  • Reliability
  • Observability

1M+

Transactions a day through the decisioning system

HowRisk variables are pre-aggregated from Kafka before a decision needs them.

  • Kafka
  • Scale
  • Low latency

0

Customer-facing downtime at go-live

HowMulti-region EKS scales out and fails over when health checks demand it.

  • AWS
  • Kubernetes
  • High availability

15+

Production releases in a year, zero defects

HowJenkins, Spinnaker, and GitLab pipelines gate releases with Sonar and Checkmarx.

  • CI/CD
  • Governance
  • Quality

15+

Systems and platforms designed and shipped

HowOrchestrators, streaming, and cloud services, each designed and then shipped.

  • Architecture
  • Microservices
  • Platform design

3+ services extracted from the WCF monolith: hotlist, negative data, and customer profile.

Selected work

The systems, and the .NET and AI inside them.

Decisioning

Legacy BRMS modernization

Led the live fraud and risk estate from FICO Blaze Advisor to Sparkling Logic. The surrounding services stayed in .NET, and an LLM path converted and checked the rules before traffic moved.

.NETSparkling LogicFICO Blaze AdvisorClaudeAWS EKS
  • Delivered 10,000+ business and engineering rules in three months, with decision parity and zero production defects.
  • Used Claude to convert rule syntax, then proved each strategy in simulation and a ramp-up so production traffic never saw an untested decision.
  • Held the integration layer in C# and ASP.NET Web API: SQL Server, Kafka, and a rule engine wired in through dependency injection, so Blaze could be swapped without a rewrite of the orchestrator.
  • Placed Sparkling Logic on Amazon EKS with multi-region scaling, Ping for identity, and Dynatrace on the platform. Fifteen production releases that year shipped with no defects.
Active-region failoverHealth checksActive regionOne flag, two sitesRegion AJobs enabledRegion BJobs pausedScheduled refresh keeps the flag current
Active-region failover. Health checks set which region is live. Jobs run there and stay paused on the standby.

AI · LLM · RAG

Rule analysis with retrieval and a model

Analysis sitting on the decisioning estate: retrieve the rules and the production behavior, then let a model compare engines and forecast a strategy before it goes live.

.NETPythonRAGClaudeSnowflake

Cloud

Risk and fraud decisioning platform

The .NET decisioning orchestrator for digital money transfer: services, Kafka, vendors, and a path that stays up when a region does not.

ASP.NET CoreKafkaCouchbaseSpring BootAWS

Open source

IntelliFlo

Open source on GitHub: a Splunk-style log platform you run on infrastructure you own, with familiar search and no Splunk license.

JavaScript.NETSPL-style searchAWS
View on GitHub

System shapes

Four patterns the platforms are built on.

Original sketches of the designs: a config plane, a two-region data path, a stream retry connector, and active-region failover.

Config plane.NET serviceJava serviceConfig serviceRuntime propertiesSource controlStream workersData services
Config plane. Services publish configuration. One service stores it in source control and fans properties out to workers.
Two regionsIngressRegion ARegion BProfileSessionAccountData clusterProfileSessionAccountData clusterasync copy
Two regions. Each region has its own ingress, profile services, and data cluster. Changes copy across when the pair is healthy.
Stream retryPublishersSuccessConnectorReads runtime configRetry storeRepublish
Stream retry. Success returns immediately. Failures sit in a retry store, then the connector republishes them.
Active-region failoverHealth checksActive regionOne flag, two sitesRegion AJobs enabledRegion BJobs pausedScheduled refresh keeps the flag current
Active-region failover. Health checks set which region is live. Jobs run there and stay paused on the standby.

Skills

Each tool, and the system it shows up in.

Expertise01–05
  • Architecture
  • Rules engines
  • API design
  • Leadership

01 / Direction

Set the end state, then build it.

Reference models, design reviews, and a rules estate the orchestrator can swap without a rewrite.

Core

  • C# / .NET

  • Architecture

  • Microservices

  • API design

  • Rules engines

Cloud and data

  • AWS

  • Kafka

  • Kubernetes

  • Docker

  • Couchbase

  • SQL Server

  • High availability

AI

  • LLM

  • RAG

  • AI

Delivery and leadership

  • Leadership

  • CI/CD

  • Observability

  • Security

Also worked with

  • Java

  • Angular

Path

Where the work happened.

Seven years of fraud and risk decisioning at Western Union, after four years building banking platforms at Zensar. Eleven years in total.

Decisioning & rules

Risk and fraud orchestrator, a highly available decision engine on FICO Blaze Advisor, and the age and velocity variable framework.

Modernization & cloud

ASP.NET Core and Spring Boot in place of legacy web and Windows services, the move from on-premises to AWS, and metadata APIs for a shared data contract.

Integration & streaming

Vendor orchestrator on Apache Camel, Spring Boot, and Couchbase. An auto-scaling Kafka consumer framework. Chargeback and fraud notices on Spring Batch.

Recent key achievements

What changed in production.

  • Rule migration at scaleArchitected, built, and delivered 10,000+ fraud and risk rules from FICO Blaze Advisor to Sparkling Logic with an AI, Claude-driven approach.
  • Sev-1 incidents to zeroBrought Sev-1 incidents from about 70 a month to zero by applying CQRS.
  • Production-traffic simulationReplays real traffic without breaking GDPR or customer privacy, so fraud strategies are tested before go-live without risking conversion.
  • Pre-aggregated variablesA platform that computes risk variables from Kafka streams before the decision needs them.
  • Auto-failoverHigh availability through an auto-failover mechanism with dynamic health-check monitoring.
  • Orphan transaction handlingCancellation and queue handling for orphaned transactions with Kafka Streams in .NET.
  • Config kept out of codeA custom Spring Cloud Config service for .NET apps, so configuration is not compiled into the service.
  • Deployment you can seeCI/CD automation with governance and deployment dashboards, so a release is visible before it is surprising.

Western Union

Aug 2019 — presentPune · Team of 10 · Mentor and delivery lead

  • .NET
  • ASP.NET Core
  • Spring Boot
  • Kafka
  • MSK
  • Couchbase
  • Docker
  • ECS
  • EKS
  • API Gateway
  • S3
  • Jenkins
  • Snowflake
  • Splunk
  • Dynatrace
  • Sparkling Logic
  • FICO Blaze Advisor
  • Join the business discussion, read the current system, and draw the end state against the enterprise architecture.
  • Deliver digital risk decisions and fraud mitigation from design through production.
  • Prove new technology with proofs of concept, including Kafka streaming and entity microservices.
  • Own the reference architecture and the reusable building blocks other teams plug in.
  • Run architecture reviews and UML for the current and end-state blueprints, including Infosec review for security and GDPR.
  • Mentor design, TDD, and code review, and stay with L2 until a production issue has a permanent fix.

  1. May 2026 — present

    Engineering Manager, Tech Lead & Senior Architect

    • Led the BRMS move of 10,000+ rules from FICO Blaze Advisor to Sparkling Logic in three months, and designed the EKS estate: capacity, network, multi-region deployment, auto-scaling, and auto-failover.
    • Lead a team of 10 across design, TDD, code review, and production fixes, and run delivery: backlog, huddles, velocity, capacity, RFPs, onboarding, and appraisals.
    • Set end-state designs against the enterprise architecture, the reference model, and plug-and-play building blocks, then review them with Infosec for security and GDPR.
    • Architected AI decisioning analysis in .NET and Python with RAG, and a Rule Evaluator on Snowflake using Claude to compare the two engines.

  2. May 2023 — Apr 2026

    Solution Architect

    • Designed a containerized, auto-scaling Kafka consumer framework on the Observer pattern, reused across business cases.
    • Led the move of .NET orchestrators, microservices, and Kafka consumers from .NET Framework to modern .NET on ECS, off Windows servers and IIS.
    • Architected API access for the decisioning APIs with Ping and Okta on API Gateway: access control, rate limits, and throttling.
    • Built cross-region auto-failover for .NET services, switching on dynamic health checks when a component or a region fails.

  3. Mar 2021 — Apr 2023

    Senior Associate

    • Used the Task Parallel Library for concurrent rule execution and result aggregation, and a custom dependency-injection container around the rule engine.
    • Diagnosed long GC pauses, Large Object Heap pressure, thread starvation, and CPU suspensions, then cut memory use with recyclable streams.
    • Isolated the leaking FICO Blaze Advisor component into a dedicated WCF service, and automated IIS pool recycle and warm-up so leaks did not take the orchestrator down.
    • Built the Kafka computation framework for age and velocity variables.

  4. Aug 2019 — Feb 2021

    Associate

    • Built a reusable authentication handler so external REST calls in the .NET orchestrator take their auth from configuration.
    • Earned Rising Star for a .NET utility that moved SOAP UI cases to Karate and cut the licensing cost of doing it by hand.
    • Moved ASMX services onto the risk and fraud .NET orchestrator, with a dynamic data model so the framework could be reused.
    • Mapped on-premises .NET and Java/Scala apps on IIS and OpenShift for a move to EKS, EC2, and ECS.

Solution highlights

Systems designed and put into production.

Technical lead

Sparkling Logic rules migration

  • Decision flows for money transfer and customer onboarding, with capacity planning and the network design on EKS.
  • Claude-driven rule syntax conversion, then simulation and ramp-up with no upstream or downstream impact.
  • Multi-region, auto-scaling, auto-failover. A Rule Evaluator on Snowflake with Claude checks that FICO Blaze Advisor and Sparkling Logic still agree.

Architect

Risk and fraud platform

  • The .NET decisioning orchestrator for risk and fraud on digital money transfer, on-premises and on AWS.
  • Rebuilt REST, WCF, and legacy ASPX in ASP.NET Core and Spring Boot, integrated with FICO Blaze Advisor.
  • Logging out to Snowflake, Splunk, Dynatrace, and Grafana. Multi-region failover from application and database health checks.
  • PCI and PII encrypted at rest with ALE KeySecure.

Technical lead

Vendor gateway

  • Vendor Gateway and the orchestrator message adapter on Apache Camel, Spring Boot, and Couchbase.
  • 15+ fraud and risk scoring vendors behind one configuration-driven framework.

Senior associate

Kafka consumers and async processing

  • A containerized, auto-scalable Kafka consumer framework on the Observer pattern for async business cases.
  • Chargeback and reported-fraud notices processed with Spring Batch.
  • Orphan transaction cancellation and queues with Kafka Streams in .NET.

Senior associate

Age and velocity variables

  • Pre-aggregated risk variables computed from Kafka and written to Couchbase.
  • Hosted on Fargate behind an application load balancer.

Zensar Technologies

SILICA

Software Developer

Oct 2015 — Jul 2019 · Pune · Team of 20

Investment products and shareholder administration across Africa, Europe, the Middle East, and Asia, with a material share of the South African market. The job was to keep a large volume of banking data accurate.

  • C#
  • .NET Core
  • Web API
  • MVC
  • WCF
  • Azure DevOps
  • AlpacaJS
  • OpenAPI
  • OAuth
  • C++
  • VB
  • PL/SQL
  • Angular
  • Three years and ten months of banking-domain delivery.
  • Moved the project to Azure DevOps independently, including the pipelines and the repositories.
  • Wrote technical designs from functional requirements and joined onsite clients for walkthroughs and status calls.

Awards and appreciation

Noted by the company and the client.

  • Rising Star

    Western Union

    A .NET utility that moved SOAP UI cases to Karate and removed the manual licensing cost.

  • 3 Quarterly Awards

    Western Union

    Awarded across three quarters.

  • Young Star

    Zensar

    Awarded at Zensar.

  • Client appreciation

    Zensar

    Client appreciation at Zensar.

B.E. Computer ScienceRGPV Bhopal · 2011–2015
AWS Solutions ArchitectAmazon Web Services · 2024
LanguagesEnglish, Hindi

For hiring teams

Open to a Tech Lead or Senior Architect role.

The current role is engineering manager, tech lead, and senior architect. Hands-on .NET and AWS, with LLM and RAG on rule migration and decision analysis. Send the role and I will reply from my inbox.