Headline

Principal Software Architect | BeneFi | Microsoft | Paytm | BlackRock

About

I am a Software Architect with 9.5+ years of experience designing scalable backend platforms, distributed systems, data-intensive applications, and cloud-native solutions across fintech, enterprise software, and financial services. My core strengths include backend engineering, system design, microservices architecture, financial workflows, performance optimization, cloud-native infrastructure, and building reliable platforms at scale. Over the years, I have worked on critical systems across payments, stock transactions, identity, notifications, compliance, productivity platforms, and business automation. I enjoy solving complex engineering problems that sit at the intersection of technology, product, and business impact. I have led cross-functional teams, collaborated closely with product managers and stakeholders, and translated ambiguous business requirements into robust, maintainable, and scalable technical solutions. I have hands-on experience with Java, Python, Spring Boot, distributed systems, databases, messaging systems, cloud infrastructure, and modern backend architecture. I am especially interested in platform engineering, fintech systems, system design, developer productivity, and building products that simplify complex workflows. Previously, I have worked with organizations including Microsoft, Paytm, and BlackRock, and I currently work as a Principal Software Architect at BeneFi Global.

Location

India

AI stack

claude code

cursor

langchain

langsmith

openai

opencode

Work experience

Software Architect at Benefi Global, Aug 2025 - Present · 1 y, United States, • Architected and implemented a unified payment orchestration platform spanning fiat, crypto, and stock-based transactions, developing the order-lifecycle state machine, pluggable provider layer, and idempotent reconciliation engine across Stripe, BitGo, APEX, and Plaid integrations. • Invented and developed a novel stock-transfer solution for non-matched accounts, reducing manual processing from 15 to 3 working days; provisional patent filed - Systems and Methods for Orchestrating Automated Cross-Title Securities Transfers Using Intermediate Accounts. • Spearheaded a backend-driven content-orchestration system enabling activity-based recommendations to be configured and shipped without application releases. • Architected an LLM-based content-moderation system that normalized raw charity-submitted data into structured GPT-4 prompts, defining a dual-prompt architecture, controlled model parameters, and structured outputs to generate consistent, reviewable classifications for human-led compliance decisions. • Designed the LLM observability and offline evaluation framework using LangSmith, defining request tracing, prompt and model-version tracking, token and latency monitoring, golden-dataset evaluation, and human-in-the-loop validation. • Designed and built a provider-agnostic notification platform supporting in-app, push, SMS, and email, capacity-modelled for up to 10 million daily notifications through horizontally scalable consumers, channel-specific retry handling, idempotent processing and pluggable delivery providers. • Architecting and implementing a multi-tenant identity and commerce platform with tenant-aware authorization, shared donor identity, isolated transactional data, and reusable checkout services across enterprise partner environments.

Senior Software Engineer at Microsoft, Nov 2019 - Aug 2025 · 5 y 9 mo, Noida, Uttar Pradesh, India, • Designed and built a rule-engine-based MSRC ticketing platform for the Windows org, automating IcM evaluation, ADO ticket creation, ownership assignment, ETA/priority determination, and status sync back to IcM. • Designed a RAG-based proof-of-concept for MSRC severity prediction, retrieving similar historical vulnerabilities from a vector database using vulnerability descriptions, affected components, and exploit proof-of-concepts, then using prompt engineered GPT-4o calls to generate severity recommendations for human review. • Led the development of Secure Install, a core Windows installation experience that activated when users attempted to install applications from external sources. The solution validated application authenticity and, when the application was available in Microsoft Store, guided the user through the trusted Store installation flow. The experience was configurable through Windows settings and required integration across Windows platform components, application security, and Microsoft Store service. • Architected the platform and phased migration strategy for onboarding Universal Apps into the Microsoft Store, defining service boundaries, integration workflows, rollout sequencing, and production-readiness requirements while technically leading seven engineers across global teams. • Led the European Union Data Boundary initiative by creating resources, migrating data, validating storage accounts and key vaults, and routing logs to EU clusters for EU customers. Migrated the s2s calls from ADAL to MSAL authentication for secure communication. • Modernized OneNote Class Notebook architecture by integrating Microsoft Graph APIs, enabling faster, cleaner feature development. • Improved production readiness - defining on-call structure, tuning thread pools, right-sizing machine counts for BTS traffic peaks, and driving high-severity incident RCAs. • Delivered top-requested educator features: Bulk Section Deletion and video / screen-recording in OneNote Class Notebook.

Software Engineer at Paytm, Jul 2018 - Nov 2019 · 1 y 4 mo, Noida Area, India, • Engineered a machine-learning feed recommendation system serving 1,000 TPS at ~150 ms average latency from user activity signals. • Built a heuristic engine moving business logic outside the release cycle, enabling rule changes without redeployment. • Implemented automated autoscaling via Ambari APIs, Blueprint, and YARN APIs to provision compute for data scientists during peak-utilization periods. • Owned end-to-end backend for the Hotel vertical, integrating Channel Managers to onboard new hotels; mentored a team of four developers.

Analyst at BlackRock, Jan 2017 - Jul 2018 · 1 y 6 mo, Gurgaon, Haryana, India, • Cut machine-hours by 70% by redesigning returns/attribution workflows to remove non-critical, memory-intensive reports. • Reduced machine-hour consumption by 60% by enhancing risk-analytics resource-scheduling algorithms. • Built a self-learning resource-allocation algorithm that assigned compute based on historical workload analysis.

Education

Master’s Degree, Computer Science, Delhi University

Bachelor’s Degree, Computer Science, Keshav Mahavidyalaya, University of Delhi