Headline

Software Engineer @ Livspace | Backend · Distributed Systems · Java | Spring Boot | Ex-Razorpay | Supply-chain | Fintech

About

Backend software engineer with 3+ years building and scaling distributed systems across fintech and supply-chain platforms. Currently at Livspace, I work on the core supply-chain infrastructure—order management, inventory, procurement, and logistics. I focus on concurrency control, event-driven processing, service modernization, database optimization, caching, and resilient asynchronous workflows that improve system scalability and reliability. Previously at Razorpay, I worked on core payments infrastructure — partner-bank integrations and payout reliability — building fault-tolerant orchestration with SAGA patterns, idempotent/exactly-once processing, and reconciliation against bank-authoritative state. Core stack: Java · Spring Boot · Microservices · MySQL · Redis · RabbitMQ · Kafka · Temporal · AWS · GCP · Open to connecting with people building large-scale backend and platform systems.

Location

Bengaluru

AI stack

cursor

opencode

Work experience

Software Development Engineer at Livspace, Jul 2025 - Present · 1 yr 2 mos, Bengaluru, Karnataka, India, Cut procurement costs by $ 1M/month by redesigning OMS order-expansion flow to allocate the actually-needed quantity, rounding each component up to its allocatable pack-size and de-duplicating shared hardware, while guarding concurrent allocation with Redisson distributed locks over MySQL. Eliminated 30 hours/week of manual effort and removed generation errors by automating delivery-challan creation through a Temporal workflow of idempotent, auto-retrying activities, while exposing it via Spring Boot REST APIs with a state-machine and optimistic locking for safe concurrent generation. Modernized core supply-chain microservices from Play to Spring Boot with zero customer-facing incidents by migrating REST endpoints through a phased rollout, using proxy-based traffic switching and Flyway migrations. Slashed p95 API latency by ∼60% on high-volume inventory and vendor services by redesigning SQL execution plans and introducing composite covering indexes that turned a ∼115-second, 27M-row full-table scan into an index-only scan, also eliminating N+1 query problem while routing read-heavy traffic through MySQL read replicas. Improved messaging throughput and reduced consumer lag across 100+ RabbitMQ event types serving 10+ downstream microservices by tuning queue configuration, consumer concurrency, and payload size. Reduced Redis connection usage by ∼45% and eliminated connection pool exhaustion during peak traffic by migrating from Jedis to the non-blocking Lettuce client, while preserving Redisson-based distributed locking with Ehcache and Memcached tiered caching.

Technical Consultant at Razorpay, Jan 2024 - Jul 2025 · 1 yr 7 mos, Bengaluru, Karnataka, India, Onboarded 100+ enterprise merchants processing 1cr+ active daily transactions with 15 TPS onto RazorpayX Payouts by integrating IP-based APIs into their finance-management systems, guaranteeing exactly-once payouts on at-least-once delivery through idempotency keys and Kafka consumer de-duplication. Architected a fault-tolerant payout orchestration layer using asynchronous messaging and SAGA patterns, isolating partner bank failures through circuit breakers, bounded retries, timeout handling, and reconciliation jobs that converged internal payout state with bank-authoritative outcomes, improving payment reliability. Built a reconciliation service with Python ETL pipelines that match the internal ledger against SFTP bank files, while tolerating late and out-of-order data through idempotent consumers. Reduced recurring payment incidents with a bank-error-classification framework that distinguished retryable from terminal failures, powering a policy-based auto-retry engine and surfacing previously unmapped failure codes. Minimized cloud-specific implementation effort by designing a cloud-agnostic storage abstraction over AWS S3 and Google Cloud Storage behind a single ports-and-adapters interface, enabling environment-level switching.

AWS Cloud Intern at ETHNUS, Aug 2021 - Oct 2021 · 3 mos, Bengaluru, Karnataka, India, Built a serverless facial-recognition platform using AWS Lambda, API Gateway, DynamoDB, and S3, enabling automated face matching through an event-driven architecture. Improved face-recognition accuracy to ∼95% by implementing cosine-similarity-based embedding comparison and optimizing similarity thresholds to balance precision and recall.

Education

Bachelor of Engineering - BE, Computer and Information Sciences, General, Dayananda sagar institutions