I build scalable Java microservices and event-driven systems that stay fast under load.
7+ years at Nagarro and TCS, promoted twice. Most recently I architected booking modules for a
nationwide automotive platform, led a team of 3, and shipped an archival pipeline that moved 300M+ records.
Backend engineer who designs for failure and ships for scale
I'm a Senior Software Engineer with 7+ years of experience building Java microservices and event-driven systems with
Spring Boot, Spring WebFlux, Apache Kafka and AWS.
I like owning a service end to end: the API contract, the service boundaries, the data model and the rollout. I build systems that stay up when
something downstream doesn't, using circuit breakers, retries with backoff, partitioned consumers and async workflows that isolate failures.
I've grown from Assistant System Engineer at TCS to Associate Staff Engineer at Nagarro, leading engineers through design and code
reviews and driving root-cause analysis for critical incidents. I hold an M.Tech in Software Systems from BITS Pilani.
Based inGurugram, India
AvailabilityImmediate joiner
FocusMicroservices, events
Last roleAssociate Staff Eng.
SandeepKumar.java
@Servicepublic classSandeepKumarimplementsSeniorEngineer {
private final int years = 7;
private finalList<String> stack = List.of(
"Java", "Spring Boot", "WebFlux",
"Kafka", "AWS");
// Requirements in, resilient services out.@OverridepublicMono<Impact> build(Requirements req) {
returndesign(req)
.flatMap(this::shipWellTestedServices)
.retryWhen(Retry.backoff(3, Duration.ofMillis(200)));
}
}
02 // experience
7+ years, two companies, two promotions
Nagarro
▲ promoted 2024Gurugram · Jul 2022 – Sep 2026
Associate Staff Engineer
Architected the Test Drive and Follow-Up modules of an enterprise automotive resale platform serving 600+ dealer outlets in 300+ cities, owning API contracts, service boundaries, data model and rollout.
Engineered reactive, non-blocking Spring WebFlux services that handle high-concurrency dealer and customer traffic with low latency.
Created a serverless archival pipeline on AWS (Lambda, SQS, DynamoDB, S3, Glue) that moves 300M+ aged records to low-cost S3 storage, replacing a manual process.
Decoupled 4 services with async SQS, SNS and Lambda workflows, so one slow consumer no longer blocks booking and follow-up flows.
Led a team of 3 engineers; set standards for microservice boundaries, API design and CI/CD, and drove root-cause analysis for critical incidents.
Spring WebFluxAWS LambdaSQS / SNSDynamoDBS3Glue
Senior Software Engineer
Developed 3 Spring Boot microservices for a large-scale e-commerce Order Management System processing high-volume, real-time orders.
Streamed order events between services through Apache Kafka with reliable, real-time delivery.
Hardened the platform for peak sale traffic with circuit breakers, retry with backoff and partitioned Kafka consumer groups.
Delivered async notification and workflow services for an automotive CRM using Spring Boot, AWS Lambda and SNS/SQS.
Spring BootApache KafkaCircuit breakerMicroservices
Tata Consultancy Services
▲ promoted 2021Gurugram · Feb 2019 – Jul 2022
System Engineer
Built and maintained Java backend services and RESTful APIs for enterprise banking applications on IBM WebSphere and JBoss.
Automated build and release pipelines with Jenkins, Git and shell scripting, cutting deployment time by 50% (40 → 20 min) and removing manual release steps.
JavaREST APIsJenkinsWebSphereJBoss
Assistant System Engineer
Implemented the TCS BaNCS capital markets platform for 4 institutional clients, supporting upgrades and 8 production releases.
Resolved release-time defects on JBoss EAP 7 and IBM DB2 through log monitoring, debugging and Bash automation.
Worked with business and QA teams on requirements, sprint planning and cross-team coordination in Agile delivery.
TCS BaNCSJBoss EAP 7IBM DB2Bash
03 // selected work
Problems I've solved, and what changed
Client work is under NDA, so these describe the problem, the design and the result without naming the products.
Data engineering · AWS serverless
300M+records archived
Serverless archival pipeline
Moving aged records out of the live store into cheaper storage used to be a manual job. I built a serverless, event-driven pipeline that
does it automatically: batches flow through SQS, Lambda and Glue into low-cost S3 storage. It has moved 300M+ records and replaced the manual process.
AWS LambdaSQSDynamoDBS3Glue
DynamoDBaged records
↓
SQSbatches buffered, retried on failure
↓
Lambda + Glueprocess and transform
↓
S3low-cost long-term storage
Architecture · Reactive
600+outlets · 300+ cities
Test Drive & Follow-Up modules
Architected two core booking modules for a nationwide automotive resale platform: API contracts, service boundaries, data model and rollout,
on reactive Spring WebFlux services built for high-concurrency dealer and customer traffic.
Spring WebFluxREST API designMicroservices
Resilience · Event-driven
4services decoupled
Failure isolation with async workflows
A single slow consumer could stall booking and follow-up flows. Moving 4 services to asynchronous SQS, SNS and Lambda workflows isolated
failures, so each flow now degrades on its own instead of taking the others down.
SNSSQSLambdaEvent-driven
Streaming · E-commerce
Peaksale traffic
Event streaming for an e-commerce OMS
Built 3 Spring Boot microservices for a high-volume Order Management System, streaming order events through Kafka, and hardened them for
sale spikes with circuit breakers, retry with backoff and partitioned consumer groups.
Apache KafkaSpring BootCircuit breaker
DevOps · Banking
40→20min deploys
Release automation for banking apps
Replaced manual release steps for enterprise banking services with Jenkins pipelines and shell automation, halving deployment time.
This work was part of a TCS Technical Excellence Award.
JenkinsGitBashCI/CD
Side project · Built and run in production
ATS Tailor: resume checker and AI tailoring
A live product I designed, built and operate on my own server. Upload a resume to get an instant ATS score, match it against a job
description, and get a tailored DOCX and PDF. Scoring is rule-based and deterministic, and basic tailoring runs on a
self-hosted LLM. It has Google sign-in, a job queue with background workers, payments and an admin dashboard.
