Manupati Ramana Kumar
Entry-Level Software Engineer with internship experience designing and building scalable applications using Type Script, Node. js, NestJS, Graph QL, and Postgre SQL. Experienced in developing REST APIs, backend workflows, caching strategies, asynchronous processing, and offline-first applications with a focus on reliability, performance, and maintainability.
Experience
SDE Intern
SLOOZE
Designed and implemented backend workflows using NestJS, Graph QL and Prisma for Goods Receipt Notes (GRN), claims and dispute management, including relational schema design and domain-specific business logic. Built an in-app semantic search prototype for GRN records using embeddings and a locally hosted LLM with Ollama to improve document retrieval while keeping data within the local environment.
Developed a Python-based web crawler using HTTPX, Beautiful Soup and Playwright to extract structured supplier information from dynamically rendered India Mart pages.
Full Stack Developer Intern
YEBSYS
Implemented resilient payment workflows with retry mechanisms and state consistency to ensure reliable real-time transaction processing. Programmed idempotent webhooks to guarantee exactly-once execution for critical financial transactions, improving system reliability.
Optimized high-traffic API endpoints by integrating Redis caching layers and asynchronous worker queues, reducing query latency by 30%.
Education
Dhanalakshmi Srinivasan University
B.Tech
CGPA: 7.69 / 10
Skills
TypeScriptJavaScriptPythonSQLReact.jsNext.jsZustandNode.jsNestJSExpress.jsREST APIsGraphQLWebSocketsPostgreMongoDBRedisIndexedDBPrisma ORMEvent-Driven SystemsOffline-First ArchitectureMulti-Tenant ArchitectureIdempotencyRAG PipelinesDockerAWS(EC2S3)NginxGitHub ActionsGit
Projects
TaskGlyph — Offline-First Workspace
Next.js, IndexedDB, PostgreSQL
Built a robust full-stack productivity application utilizing a local-first architecture to allow seamless offline user interactions. Developed a custom synchronization engine to reconcile client-server states, handling out-of-order updates and resolving conflicts using Last-Write-Wins logic.
Integrated exponential backoff strategies to achieve 96% reliable data synchronization across devices under unstable network conditions.
TriaGen — Incident Management System
NestJS, Redis, AWS
Constructed an event-driven backend using Web Sockets to broadcast low-latency, real-time updates to distributed users during critical outages. Configured asynchronous background job queues via Redis for automated SLA monitoring and retry mechanisms, decoupling heavy operations from standard requests.
Streamlined asset management by offloading file uploads to AWS S3 using pre-signed URLs, minimizing server bandwidth and improving response times.
Rephrase AI — Distributed Processing Pipeline
Node.js, Redis, LLM APIs
Developed a multi-stage ingestion and chunking pipeline capable of processing large-scale text documents while navigating strict AI API token limits. Built automatic API failover across multiple LLM providers to maintain service availability during outages.
Eliminated redundant external API calls by implementing a Redis caching layer, significantly reducing operational costs and accelerating system response times.