Impact
$500K+
Annual savings
99.9%
System uptime
9+
Products shipped
60%
Fewer manual workflows
Experience
Microsoft
2022 - PresentSenior Software Engineer · Time Agent, Expense Agent, Workflow Orchestration
“Architected and delivered Time Agent on Copilot Studio from scratch, leading a team of 5 to ship 9+ products to GA.”
Architected and delivered Time Agent on Copilot Studio from scratch: owned technical design, led a team of 5 engineers, and shipped 9+ products to GA for enterprise customers.
Contributed to Expense Agent - started as a hackathon prototype, adopted as a company-wide product. Monitors shared mailboxes, parses receipt emails with Azure OpenAI, and auto-creates Expense Lines and Reports in Dynamics 365 F&O.
Delivered Project Budgeting feature in Dynamics 365 Project Operations as part of a team of 4: built budget management workflows with time-phased forecasting, cost tracking, and approval workflows.
Built prompt pipelines and inference infrastructure (retries, token budgets, model routing) for Azure OpenAI and SLM integrations, keeping LLM-powered features reliable at scale.
Designed and built a workflow orchestration platform with C# microservices, REST APIs, and a React/TypeScript frontend, reducing manual workflows by 60%.
Re-architected execution layer consolidating 10+ legacy pipelines into a unified system: 40% cost reduction ($500K+ annual savings), 30% faster incident resolution.
Drove technical alignment across 4 orgs (Product, Platform, Expense, Sales); unblocked critical dependencies and accelerated GA timeline by 2 months.
Designed integration platform for external partners: defined OpenAPI contracts, auth flows, versioned schemas, and webhook patterns; onboarded 4+ partner teams to production.
Built CI/CD infrastructure with Azure DevOps; implemented telemetry (Application Insights) enabling P50/P99 latency monitoring, A/B experiments, and 25% faster defect detection.
On-call DRI for production systems; led incident response, conducted post-mortems, and improved system uptime from 99.5% to 99.9%.
Mentored engineers on Copilot Studio and Azure OpenAI; ran knowledge-sharing sessions and conducted technical interviews.
Oracle
2019 - 2021Software Engineer · Healthcare Supply Chain
“50K+ daily transactions at P99 latency of 300ms for Fortune 500 healthcare clients.”
Owned end-to-end delivery of Recall Management system (Healthcare Supply Chain) from design to production; processed compliance-critical transactions for Fortune 500 healthcare clients.
Designed and implemented high-throughput REST APIs (Java, C++) handling 50K+ daily transactions at P99 latency of 300ms; optimized SQL queries reducing DB load by 35%.
Architected Redis caching layer for hot-path queries, reducing response times by 40% across distributed healthcare systems.
Built 4 production UIs (React, Oracle JET) improving user task completion rate by 25%; led UX reviews with product team.
Security Point of Contact: led threat modeling sessions, coordinated pen-testing, and remediated 15+ vulnerabilities pre-launch.
Mentored 3 junior engineers; established code review standards and conducted 20+ technical interviews.
Tarana Wireless
2018Software Engineering Intern · 5G Spectrum
Built CBRS Spectrum Access Server module in C++ for 5G spectrum allocation; achieved sub-100ms latency across 100+ concurrent device connections.
Tech Stack
Languages
AI/ML
Backend
Frontend
Data
Infrastructure
Practices
Projects
RAG Document Assistant
Built a RAG-based document assistant using LangChain and Azure OpenAI; implemented semantic search with embeddings, prompt chaining, and guardrails; exposed via FastAPI with streaming responses.
Added web-grounded answering mode with runtime retrieval and citation linking.
Built evaluation pipeline measuring relevance, faithfulness, and coherence to iterate on answer quality.
Task Management Service
Full-stack task management app with Go backend serving GraphQL APIs (queries, mutations, subscriptions), React frontend, Postgres for persistence, Redis for caching; containerized with Docker Compose.
Webhook notifications for state changes with retry logic.
Application Layer DDoS Detection
ML-based detection system for slow-rate DDoS attacks; achieved 95% accuracy with <50ms detection latency; research funded by Indian government (CERT-In, MeitY).
Achievements
Delivered $500K+ annual cost savings through unified orchestration layer while shipping 9+ products on schedule.
Expense Agent hackathon project adopted as a production product used across Microsoft.
Improved system reliability from 99.5% to 99.9% uptime; reduced MTTR by 30%.
Reduced production bugs by 40% in preview releases through TDD practices and automated testing.
GATE 2017 All India Rank 892 (Computer Science) out of 100K+ candidates.
Education
National Institute of Technology
M.Tech, Computer Science & Information Security
Government College of Engineering
B.Tech, Information Technology
Best Student Award (TCS)