S.H.I.E.L.D. ARCHIVE LEVEL 7
I’m a Software Engineer II at Microsoft, where I build intelligent recommendation and workflow systems across Bing, Edge Search, and Microsoft 365 Copilot. My work spans real-time intent classification, agent discovery, multilingual prompt generation, and large-scale data pipelines.
Before Microsoft, I architected Ellucian’s production GenAI platform on AWS and improved RAG answer relevance by 60%. My research in AI and cybersecurity has led to eight publications and Virginia Tech’s 2024 Best MS Research Award.
GPA: 3.77
Worked with Dr. Jin-Hee Cho on an Interdependent Mission Impact Assessment (iMIA) framework using Bayesian reasoning to infer the performance and game-theoretic enemy behavior modeling to simulate the attack-defense interactions with the mission system.
Received the Best MS Research Award, Department of Computer Science, Virginia Tech, 2024.
CGPA: 8.93 (B.Tech: 8.67 & M.Tech: 9.70).
Worked with Prof. O. P. Vyas on my master's thesis titled "A Multi-Agent Framework to Detect In-Progress False Data Injection Attacks for Smart Grid".
Ellucian IQ GenAI: Architected and deployed a scalable, production-grade Generative Agentic AI platform using AWS Serverless (Lambda, Step Functions, DynamoDB) and AWS Bedrock, laying the foundation for company-wide AI integration.
Implemented advanced Retrieval-Augmented Generation (RAG) using AWS OpenSearch vector search, improving answer relevance by 60% through contextual grounding and semantic enrichments.
Reduced perceived latency from 10s to under 1s through asynchronous processing with SNS streaming and WebSockets, significantly enhancing user experience.
Delivered an AI-powered rich text editor in React.js, using the GenAI platform for dynamic text enhancements, context-aware content generation, and personalized AI interactions with support for rich text formats like quill delta.
Employed Test-Driven Development (TDD) and authored automation tests using Playwright, ensuring reliable high-performance backend services and UI, which reduced bugs and improved software quality.
Built a robust course selection shopping cart service using AWS Lambda and DynamoDB with a React.js frontend, scaling to handle over 10,000 peak concurrent users (PCU) during registration periods. Developed WebSocket APIs and SQS queues for real-time pricing and availability updates via DynamoDB streams, eliminating registration conflicts and improving user experience.
Migrated a monolithic patient monitoring application to a microservices architecture using Java SpringBoot, Spring Cloud, and Hibernate, improving system scalability and maintainability.
Implemented a real-time analytics dashboard using Kafka streams and Elasticsearch, enabling clinicians to monitor critical patient metrics across multiple units, resulting in 35% faster response times to clinical deterioration events.
Developed comprehensive JUnit and Mockito test suites integrated with the Jenkins CI/CD pipeline with Snyk and SonarQube scans, reducing regression testing time by 50% while expanding test coverage by 90%.
Developed a novel approach using GANs for NURD correction in the intravascular ultrasound images using PyTorch and received the Philips spot Award for my contribution.
Developed a comprehensive framework that scans the devices and detects malicious activity in an intranet in real-time for SCADA/ICS networks using Snort, Scapy, Flask, D3.js for data visualization.
Reliable distributed services and data pipelines built for production scale.
Serverless and event-driven systems across Azure and AWS.
LLM agents, RAG, recommendations, evaluation, and intelligent workflows.
Research-backed mission resilience, network security, and adversarial modeling.