Software Engineer — Technology Consulting - EY (Ernst & Young) - Remote, India
(2025-10)
- DPG FPO Platform (Live: api.fposhaktidpg.in): Architected and built a greenfield microservices platform for Farmer Producer Organizations from scratch using Java Spring Boot, with domain-driven service boundaries and PostgreSQL (database-per-service) managed via Flyway migrations. Platform is live and serving 20,000+ registered FPO users.
- gRPC Service Mesh: Designed gRPC inter-service communication with parallel fan-out calls for cross-service data aggregation, backed by Resilience4j circuit breakers, retries, and timeouts — replacing REST-over-HTTP for internal calls and cutting inter-service latency.
- Zero-Loss Event Publishing: Implemented the transactional outbox pattern with RabbitMQ for guaranteed at-least-once domain-event delivery, powering reliable CQRS read-model projections with zero message loss across service crashes or broker outages.
- Authentication & Authorization: Integrated Keycloak for centralized OAuth2/OIDC authentication with fine-grained role-based access control across all services; applied the Strategy pattern to isolate business rules per FPO scheme, enabling new variants without modifying core service logic.
- Tech Stack: Java, Spring Boot, gRPC, PostgreSQL, RabbitMQ, Keycloak, Resilience4j, Flyway, Docker
Software Engineer - Markopolo AI - Remote
(2022-08 - 2025-10)
- First-Party Tracking Pixel (Marktag): Built a server-side, first-party tracking solution in JavaScript that bypassed third-party cookie and browser restrictions for continuous user identification across ad campaigns, improving web tracking accuracy by 40% and reducing data loss by 35%. Ingested 30–50 clickstream events/sec via Kafka producers/consumers, decoupling event capture from storage writes and keeping p99 ingestion latency stable under burst traffic.
- Write Amplification Elimination: Replaced a fan-out write pattern that simultaneously wrote each captured event across 6 MongoDB collections (master, day, month, product, purchase, cohort) with a single master-collection write plus a 24-hour cron aggregation pipeline that batch-inserts pre-aggregated rollups. Reduced total MongoDB R/W operations by 75% and cut reporting dashboard load times by 30%.
- Analytics Dashboard Backend: Built scalable backend services for sales and campaign analytics consolidating data from 5+ ad platforms (Google, Amazon, Facebook, Shopify, HubSpot, WooCommerce) with optimized schema design, cursor-based pagination, and response times under 200ms using Redis caching.
- Bulk Email & Messaging System: Developed a bulk email journey/automation system on AWS SES + SNS + SQS handling 100,000+ users/month with full throttling control, real-time open/click tracking, and delivery optimization.
Scaled
RabbitMQ-based workflow automation from 10K to 50K+ messages/day with zero data loss.
- Shopify App Integration: Built Shopify store app with OAuth2 authentication, fetching customer and product data via REST API and GraphQL, and managing scalable webhook event ingestion from Shopify and WooCommerce with data validation and integrity guarantees.
- Data Science Support: Ran clustering and cohort analysis over millions of ad-event records using PySpark, PostgreSQL, and BigQuery to power audience segmentation decisions.
- Tech Stack: Python, Flask, Node.js, TypeScript, PostgreSQL, MongoDB, Kafka, RabbitMQ, Redis, AWS, GCP, CI/CD
