Software Engineer at Microsoft

I build software that turns complex ideas into useful products.

I like working on messy, high-stakes problems and turning them into products people can trust. At Microsoft Security, I build AI-assisted investigation experiences; beyond work, I prototype tools across finance, developer productivity, and applied AI.

Explore my work ↓ Let’s talk

GitHub LinkedIn

C# / .NET✦React & Next.js✦Azure & AWS✦Python & FastAPI✦RAG & AI Agents✦SQL & Data

02 / About

I started as a full-stack engineer, working across every layer of a product: React interfaces, .NET APIs, databases, cloud infrastructure, testing, and delivery. That end-to-end foundation taught me how useful, dependable software is designed, built, and operated.

As the products and problems grew, my focus expanded from shipping complete applications to making those applications more intelligent. NYU and Alpheva AI moved my work toward mobile products, RAG, financial workflows, and coordinated AI agents.

That path brought systems engineering and product thinking together. Now at Microsoft Security, I build AI-assisted investigation experiences grounded in evidence, risk, and reliable services. The goal is AI that remains useful, explainable, and worthy of trust.

03 / Experience

Engineering with measurable impact.

Jan 2026 — Present

Seattle, USA

Microsoft

Software Engineer

Built full-stack AI-powered investigation workflows for Microsoft Security’s Data Security Investigations team, developing customer-facing interfaces and back-end services to analyze Microsoft 365 content such as emails, Teams messages, documents, and investigation datasets, improving investigation efficiency by 40%.
Designed semantic retrieval capabilities using embeddings and vector search for Security Investigations, enabling analysts to query investigation-scoped Microsoft 365 content using natural language and surface related evidence even when exact keywords were missing, increasing relevant evidence discovery by 30%.
Developed LLM-based AI analysis workflows for Data Security Investigations to categorize sensitive content, evaluate exposure severity, assign risk scores, and generate mitigation insights, improving investigation prioritization accuracy by 25% and reducing manual review effort.

React.NET / C#AzureRAGVector Search

Aug 2025 — Dec 2025

New York, USA

Alpheva AI

Software Engineer

Deployed an iOS and Android app with Plaid-based bank account linking and personalized dashboards, increasing successful account connections by 30% and cutting setup time by 40%.
Created an AI-agent-powered financial advisory platform integrating retrieval-augmented generation, contextual memory, and risk-management agents, enabling personalized investment insights for 1,000+ simulated user portfolios.
Enhanced multi-agent orchestration across Recommendation, Portfolio Optimizer, and Risk Management agents using Python, Redis cache, and vector databases, improving response accuracy by 35% and reducing query latency by 40%.
Built and normalized financial data models for users, portfolios, transactions, loans, and tax records with automated validation and monitoring, ensuring data consistency across 8+ services and supporting scalable API endpoints.

PythonReact NativeAWSMCPPinecone

Feb 2022 — Aug 2023

Bangalore, India

Zensar Technologies Ltd

Software Engineer

Engineered a microservices e-commerce platform with ASP.NET Core, designing and implementing robust REST APIs and utilizing AWS ECS, cutting infrastructure costs by 10% and employing AWS Lambda for serverless ETL to enable efficient data integration.
Optimized database performance with Amazon RDS for PostgreSQL, query optimization, and Amazon ElastiCache for Redis caching, and applied AWS Elastic Load Balancer for traffic distribution, achieving a 30% improvement in response times.
Managed development workflows using Git and Docker, facilitating efficient CI/CD processes and integrating NUnit and Selenium testing, reducing deployment times by 25%.
Accelerated Agile Scrum delivery by leading sprint planning, code reviews, and mentoring junior developers, achieving a 15% code-quality improvement through peer programming and linting standards.
Collaborated with DevOps, QA, Product, and other cross-functional teams, accelerating product development lifecycles by 40% and ensuring seamless integration from ideation to production.
Implemented a Kafka-driven streaming data pipeline that ingested and processed records from an AWS DynamoDB repository. Developed primarily using Java, it enabled critical pattern analysis that improved downstream reporting accuracy by 50%.

.NETReactAWSKafkaDocker

04 / Selected work