Tushar Kumar

• Developed a high-performance distributed data write engine to process terabyte-scale datasets, handling 5+ million daily write-intensive requests with low latency and high reliability. • Identified inefficiencies in a distributed service responsible for rewriting Parquet files across data lake destinations. Designed and implemented multi-tenant request handling, reducing AWS cloud costs by 40% while maintaining performance. • Designed a memory-aware request acceptance algorithm that uses real-time load prediction to make fast accept/reject decisions, ensuring optimal resource utilisation under dynamic load conditions. • Optimised cloud auto-scaling by tuning scaling logic based on concurrency levels, server load percentages, and pod idle times improving resource usage across regions. • Developed an algorithm to maintain eventual consistency between Delta Lake and Apache Iceberg table format, ensuring data integrity in the event of catastrophic failures. • Built a scalable garbage collection system for cloud storage that safely removed orphaned files, cutting storage costs by up to 80% without compromising data integrity. • Introduced a range-based conflict detection mechanism using Iceberg's internal APIs to ensure consistent writes in high-concurrency environments. • Designed and implemented cross-catalog synchronisation for a table across multiple metadata systems (e.g., Glue, Apache Polaris), with an integrated rollback framework to ensure consistency and fault tolerance. • Consistently earned top-tier performance ratings (“Often Exceeds Expectations" or “Exceptional Performance”) in 80% of company reviews, reflecting strong contributions and reliability over time.

-Worked on Designing and implementing the Approval Framework of the application enabling any critical operation to go through centralized Approval Pipeline. -Designed mission critical Approval APIs enabling Reject and Approve functionalities through REST API calls.

-Worked with Externalizing the application configuration ensuring a central place to manage external properties for applications across all environments, enabling run-time adaption of critical Application properties without redeployment activity. -Have experience and knowledge of working on RESTful Web Services. Have Written several REST APIs for global financial solutions ensuring low latency by writing concise and efficient code. -Worked with the Service Discovery and Registry using Eureka thereby enabling microservices to interact with dynamically assigned network locations. -Have worked on several Java 8 features such as CompletableFuture, Functional Interface, Lambda Expressions, Stream API, Collectors class and Optional to adapt to the latest coding standards of the language. -Have knowledge of OOPS Design Patterns and Design Principles, worked on Creational and Structural Design Patterns such as Builder, Adapter, Decorator to build concise and efficient solutions to promote re-usability as well as extensible solutions. -Have written complex batches to compute critical information required in the Transaction modules of Financial Application. -Have designed the Validation framework, of the Application using Generics to enable across-application usage to create Validation pipelines, ensuring the execution of data and Business Validation. -Have worked on the Swagger Documentation of the REST APIs in the project enabling the Documentation to showcase various information like Request & Response Body, Mandatory Request Parameters etc required for the end-users