Kaushik

Senior Software Engineer

About

I care about two things: systems that don't break in embarrassing ways, and codebases that the next engineer doesn't curse. Everything else — cloud, microservices, observability, AI tooling — is just the current vocabulary for those two problems.

Professional Experience

Globant

Senior Software Engineer

August 2024 – Present
Ahmedabad (Remote)
  • Led OpenTelemetry adoption across distributed services, enabling end-to-end distributed tracing, unified metrics, and structured logs — replacing fragmented Azure Table Storage logging with a production-grade observability stack.
  • Optimized a high-volume Job Tracker handling 100K+ rows by rewriting stored procedures, adding indexes, and implementing archival jobs — reducing average response time from 60s to under 1s.
  • Re-architected a CSV export API for large financial datasets using streaming responses, rented memory pools, and read-only spans — doubling concurrent user capacity and increasing throughput from 3k to 7k req/s under load tests.
  • Eliminated memory and server overhead for large file downloads (GB-scale) by replacing in-memory fetch with signed Azure Data Lake URLs, enabling browser-native downloads resilient to page refreshes.
  • Engineered a unified NuGet caching library adopted across all services — standardizing resilience patterns, implementing proactive cache warming to prevent thundering herd, and enabling project-level cache invalidation.

Liquidly

Technical Lead

April 2020 – August 2024
Ahmedabad (Remote)
  • Architected a centralized authentication platform using IdentityServer4 and OAuth 2.0, unifying identity across multiple trading workflow applications and eliminating redundant onboarding for Organizations and Users.
  • Led migration from monolithic to distributed architecture — decomposing into Business APIs, Identity, Notification, and Background Job services — achieving 50%+ reduction in average API response time.
  • Designed and built production-grade REST APIs following Clean Architecture and TDD; implemented HTTP client code generation, saving tens of dev hours and reducing PR comments by 90% through tooling and guidelines.
  • Led the team of 8 engineers — translating business requirements into sprint tasks, conducting design reviews, driving agile ceremonies, and mentoring junior developers on architectural best practices.
  • Reduced CI/CD pipeline runtime by optimizing configuration files, cutting execution time from 15 to 8 minutes.

Marg Sahayak, Government of Gujarat

Software Engineer

April 2018 – March 2020
Ahmedabad
  • Designed and launched a scalable grievance management portal serving 1,400 government officers across 33 districts with GIS-based road lookup — handling 18,000 complaints in 3 days on a Node.js and MongoDB backend.
  • Built a stateless backend using JWT and distributed caching to enable horizontal scaling, achieving 2x throughput on existing infrastructure — replacing a slow hierarchical offline system presented to the World Bank.

Projects

AsyncEndpoints

  • An AOT-compatible distributed background job processing NuGet library supporting .NET 8–10, with Redis-based job store using Lua scripting for atomic operations and sub-millisecond API response times.
  • Engineered a resilient async processing pipeline with configurable retry strategies, bounded concurrency controls, automatic stuck job recovery, and built-in structured metrics and tracing for full observability.

Technical Skills

System Design

MicroservicesDistributed SystemsEvent-Driven ArchitectureClean ArchitectureAsync Processing

Languages & Frameworks

C#.NET CoreEF CorexUnitSQLT-SQLJavaScriptTypeScriptNode.js

Databases

SQL ServerRedisSQLiteMongoDB

Cloud & Infrastructure

DockerNGINXAzure App ServicesAzure FunctionsAzure Event GridAzure Blob StorageAzure Application InsightsAzure DevOps

Education

Vishwakarma Government Engineering College

Bachelor of Engineering in Information Technology

August 2016 – September 2020
Ahmedabad