Part of the Industry Solutions Engineering organization that works with Microsoft’s strategic partners and customers to code with their engineers to help unlock technical capabilities and pioneer novel solutions using PaaS services on Azure cloud. Emphasize on security by design in software development. Lead the team to make security the priority rather than an afterthought.

Financial Services Agentic Workflow - Designed an advanced agentic AI workflow designed specifically for the financial industry. The system speeds up response to the rapid developing market, gives insights to portfolio holdings. Leveraging the Model Context Protocol (MCP) alongside Microsoft’s AI Foundry, secured with Entra ID and content safety service, the system aims to provide an adaptable, extensible and efficient platform for rapid experimentation and iteration.

Industrial Systems Asset Knowledge Base - Designed and implemented AI copilot for Honeywell APM copilot using RAG pattern and Azure AI search. Set up evaluation platform and metrics to evaluate performance of different LLMs, embeddings, and vector stores. The system can quickly identify asset dependencies, pinpoint asset trouble spot and reduces expected time for resolution from days to minutes with a focus on grounded, safe responses. Created security plan and threat modeling for both dev loop and user scenarios. Identified and implemented mitigations. Created cost model for experimentation and production based on model selection and experimentations run for cost control.

Software Defined Vehicle E2E Framework & Control Plane – Designed and implemented highly scalable, highly available, and highly performant secure end to end vehicle and device communication platform on Azure Cloud. Created stamped architecture and Entra ID cross tenant solutions to solve a variety of use cases. Implemented control plane for the system using helath modeling based on SLA and SLOs.

Reviewed existing ZellePay system solution. Resolved efficiency issues for both batch processing and API accesses. For batch job processing, optimized Kafka consumer logic to decouple stream consumption and event processing. Reduced queue rebalancing errors and published technical blog to share with other teams as best practice. Identified redundancy in producer creation and improved message production efficiency as well as reduced producer compute usage by 90%. API response time for P99 was improved to less than 500ms. Implemented timestamped and custom avro serialization for efficient data streaming – reduce storage (long vs. string) as well as flexible formatting options.

Implemented consumer Kafka pattern to provide utility methods for teams to handle enqueue/dequeue from kafka so that each team don’t have to write their own queue client or host their own listeners and can focus on implementing business logic. Designed resilient and decoupled processors to mitigate transient errors while providing logical breaks for additional development and processing.

Modernize & Redesign client server system for enterprise employee benefit outsourcing systems to distributed systems for better scalability. Split front end UI rendering completely separate from backend business processing. The same back end system serve Web front end, VRU front end, Content Manager front end, Case management front end as well as desktop and web front end.

Desktop and web front end handles different personas with Org level security control that each persona gets their own menu and GUI and access rights to ensure data integrity and compliance.

The upgrade removed the tedious tasks to set up access points for each type of front end application.

The org level security control allows customer admins to easily configure user access without repetitive development requirements. The whole front end is completely configurable with XML configuration. Customer onboarding can be completed without dev team involvement for customer admins with only XML and general system knowledge.

• Analyze, design, and implement scalable applications for employee benefit outsourcing

• Implement Java Swing GUI in HTML5/JavaScript using dataTables (identified and fixed a performance issue when many fields are on the editor screen), jQuery, Ajax, CSS.

• Assign tasks based on team member ability to complete project efficiently.

• Mentor junior members of the team

• Evaluate and implement java reporting to replace Crystal Report that requires VBA JNI access.

• Help address client connectivity issues by reviewing network captures

Analyze, design and implement scalable enterprise applications for employee benefit outsourcing

Implement Rabbit MQ solution to decouple business logic to make enterprise application more scalable and reliable

Profile and implement performance enhancement for Java Swing App

Implement various SOAP WebServices for backend access

* Implement Participant Web using XML, XSL, HTML, CSS.

* Create bread crumb framework and workflow solution

* Use Java Interface to implement java thread pool for backend calculation processing

* Implement backend interface/middleware for WorkStream documentation

* Implement multithreaded random access client app for performance profiling