01
Design for failure
I think about timeouts, retries, idempotency, observability, and graceful degradation before a system reaches production.
Software engineer · backend & distributed systems
I build backend systems that stay understandable under load.
I'm Shiwang Kumar Rai, a software engineer focused on reliable APIs, microservices, concurrency, and event-driven systems using Java, C++, gRPC, and Apache Kafka.
350+
DSA problems solved
12+
backend endpoints delivered
5+
production modules integrated
< 4h
release turnaround after automation
Technical stack
Tools I use to turn system designs into working software.
Java
C++
Spring Boot
Spring Security
gRPC
Apache Kafka
PostgreSQL
Docker
AWS
Selected work
Projects with an engineering story.
01Read case study →
FastAPI Microservice Platform
A production-ready backend platform built with FastAPI, featuring 12+ REST endpoints, Pydantic validation, PostgreSQL query optimization, and automated CI/CD — reducing release turnaround from 2 days to under 4 hours.
FastAPIPythonPostgreSQLMongoDB
02Read case study →
Distributed Auth Service with gRPC
A stateless, role-based authentication and authorization service built with Java and Spring Security over gRPC — designed for low-latency token validation at scale across multiple microservices.
JavaSpring BootSpring SecuritygRPC
03Read case study →
Event-Driven Order Pipeline with Apache Kafka
An event-driven microservices pipeline using Apache Kafka as the backbone — decoupling order ingestion, inventory updates, and notification dispatch into independently scalable consumers.
JavaSpring BootApache KafkaPostgreSQL
How I work
Engineering decisions I care about.
02
Make the boundary clear
Good contracts reduce integration cost. APIs, schemas, events, and authorization rules should be explicit and easy to reason about.
03
Optimize with evidence
Performance work starts with measurements. I prefer profiling, query analysis, and focused changes over premature optimization.
Let's build something useful
Hiring for backend engineering, or solving a hard systems problem?
Send me the problem, constraints, or role. I'd rather start with the engineering challenge than a generic pitch.