Sekhar Ghosh

Client: BlackHawk Networks

• Led end-to-end Designed and delivered a multi-tenant Catalog-SDK supporting 100K+ tenants and
  millions of catalog/Product lookups, providing a standardized, high-performance catalog access layer
  for multiple applications. Leveraged FlatBuffers, Redis, and multi-level caching to achieve sub-second
  response times while reducing backend dependency calls and infrastructure overhead.
  ● Key Results:
  ● 2M+ requests/day.
  ● P95 latency reduced by 60%.
  ● Cache hit ratio 95%+.
  ● Throughput increased to 2.1K+ requests/sec.
  ● 15+ Consuming applications.
  ● Established AI-powered observability with New Relic to monitor JVM health, CPU, memory, GC
  behavior, and transaction performance, enabling proactive issue detection and improved platform
  reliability.
  ● Built event-driven microservices leveraging asynchronous messaging to support real-time catalog
  propagation, product lifecycle updates.
  ● Working on a cloud-native AI-powered Product Recommendation Engine on AWS leveraging Amazon
  Bedrock, OpenSearch Vector Search, and event-driven microservices to generate personalized product
  recommendations using customer purchase history, browsing behaviour, and semantic product
  matching, driving higher customer engagement and improving product discovery across a multi-tenant
  platform.

Client: OCBC Bank

• Modernised the Treasury Static Data System (TSDS) by decomposing legacy monolithic applications
  into scalable microservices, improving maintainability, deployment agility, and system reliability.
  ● Automated treasury data ingestion, validation, transformation, and distribution workflows using
  Kafka-driven microservices and Amazon S3, eliminating manual processing steps, reducing
  operational costs by 30%, and improving data accuracy and processing efficiency.
  ● Enhanced production stability through proactive monitoring, centralized logging, automated alerting,
  and root-cause analysis, reducing incident resolution time and improving overall system reliability.

Client: J.P. Morgan Chase

• Led a team of 4 engineers responsible for production support and operational stability of the Paperless
  Subscription Notifications (PSN) platform, a business-critical customer communication system serving
  5M+ customers across Chase banking products.
  ● Contributed to the design and development of the Kafka-based PSN Adapter, enabling seamless
  integration between legacy monolithic systems and Chase Cloud infrastructure through distributed
  event-driven architectures.
  Impact: Ensured timely, secure, and reliable delivery of millions of customer notifications related
  to account activity, payments, transactions, and security events, enhancing customer experience, trust,
  and engagement with digital banking services.

• Engineered Java-based APIs and microservices for document ingestion, training data management,
  and AI model integration, enabling automated extraction of structured data from unstructured
  documents.
• Built scalable cloud-native services deployed on Docker and Kubernetes, significantly
  reducing manual processing effort and improving operational efficiency.

Client: Nokia

• Developed scalable core software solutions using Java and relational databases, enhancing telecom
  system performance and cross-network infrastructure.
  ● Identified, troubleshot, and resolved critical production bugs, driving long-term reliability enhancements
  across the local code base.