Aftab Ahmed Khan

Project: CDC(Customer Data Cloud) Agentic AI Platform - Intuit

Built self-healing agentic AI system for CDC (Customer Data Cloud) pipeline onboarding, reducing manual on-call intervention.

Designed and built a RAG-powered self-healing agent using LangGraph + GenOS that auto-detects pipeline version mismatches during PROD promotion, retrieves fix steps from a knowledge base, and auto-remediates via QAL/E2E promotion — eliminating manual on-call fixes for onboarding failures.

Developed self-healing agentic AI system for CDC pipeline onboarding, minimising manual on-call intervention.

Enhanced CDC's MCP Server and Capability Plugin for natural-language attribute discovery across CDC 1.0/2.0 and UCP, integrating with GenOS Agent Registry and eval framework (GenOSEval/Langfuse) to improve routing accuracy and answer-quality monitoring.

• Created data-access tools (CdcUserTool, CdcCompanyTool) using cdcdatalib, secured by attribute allowlist and Agent Identity for service-to-service authorisation.
• CDC is a distributed, high throughput, low latency platform that provides a unified view for Cross-BU attributes for Company, User and Visitor to its clients. Existing capability in CDC allows delivery of these attributes through GraphQL APIs and real-time notification over event bus.

Tech Stack and Tools: LangGraph, GenOS (llm-execution-svc, Agent Registry, GenOSEval), RAG (rag-ingestion-svc/rag-search-svc, ChromaDB), Anthropic Claude models, CDC MCP Server, Agent Identity/authz, Langfuse.

Technologies used include Kotlin, Java 21, Spring Boot, Kafka, and Microservices.

Tools used include Splunk, Jenkins, and Argo CD.

• Worked on feature flag-related code cleanup.
• Worked on functional testing via the Karate framework.
• Worked on performance testing via Gatling.
• CAMS allowed collaboration between accounts so that a user in one account can access resources in another account for purposes such as sharing data, documents, and establishing the relation between them.
• Resolved Jira tickets with comprehensive bug fixes.
• Project: Cross Account Management Service (CAMS - QuickBooks) - Intuit.

Project: Company Creation Service (QuickBooks) - Intuit

The Company Creation Service provides the end user with the account sign-up facility in Intuit, where the user and accountant can maintain firm, client, and client books.

• Migrated REST service to GraphQL to enhance data retrieval efficiency.
• Performed KIAM to IRSA AWS-related migration.
• Implemented feature flagging to manage application features while migrating.
• Executed version upgrade for the application to improve performance.
• Resolved security defects to strengthen application integrity and compliance.
• Provided production support, executing bug fixes, managing deployments, and conducting root cause analysis.
• Worked on creating multiple Splunk dashboards for monitoring application, and API status.

Project: ITR (ItPnL Reconciliation) - Morgan Stanley.

ITRs are used to facilitate the reconciliation of the profit and loss for the overall transactions made via the ledger account.

• Implemented a new microservice component.
• Implemented a web application in Angular as client code.
• Worked on various microservice components.
• Implemented the JUnit test cases with Mockito.
• Implemented the business layer and repository layer based on the requirements.
• Documented the Spike and Enabler-related matters.

Project: MSTMS-NETS
This MSTMS (Merchant Services and Terminal Management System) is used to manage the terminals, and merchants.

• Implemented new microservice component.
• Implemented web design related to Angular 8.
• Coordinate with team members, and work on various micro services.
• Implemented the JUnit test cases with Mockito.
• Implemented the business layer and repository layer based on the requirement.
• Implemented the framework for auditing the field-level details and transactions.
• Implemented Spring Boot Admin to send notifications.
• Documenting the Spike- and Enabler-related things.
• Fixing bugs and API performance-related work, design sequence diagram.
• Worked on Pre-Prod and Production deployment in a Linux environment.
• Implemented logic for supporting Master and Visa card-related parameters.

Project: BPCL Loyalty Program
This BPCL provides an automated system for the automation of the workflow of a petrol pump. The main objective of the BPCL Loyalty Programme is to simplify and expedite the process by offering a platform where individuals can enter or update their basic information to register themselves, and to use other functionalities available in portals. With this, the entire process becomes fast and efficient.

• Implemented Spring MVC with Spring Boot.
• Implemented web design-related implementation using Angular 8.
• Implemented background process using Spring Batch.
• Implemented generating new BPCL card and related CVV.
• Implemented business logic for NACH and NEFT.
• Implemented business logic for reconciliation.
• Configured the new HSBC-related API for transactions.
• Completed Spike and Enabler tasks, as well as fixing the bug.

Project: SHELL Loyalty Programme
This SHELL Loyalty Program provides an automated system for the automation of the workflow of a petrol pump. The main objective of the SHELL Loyalty Programme is to simplify and expedite the process by offering a platform where individuals can enter or update their basic information to register themselves, and to use other functionalities available in the portals. With this, the entire process becomes fast and efficient.

• Implemented Spring MVC flow at the backend side.
• Implemented background processes using Spring Boot Batch.
• Implemented client code in C#.
• Implemented business logic for reconciliation.
• Pre-prod deployment and fixing bugs.
• Production support.