SDE 2 at Nium, open to backend roles

UtkarshKumar

Backend engineer with 2+ years building event-driven systems for fintech. I design microservices that move millions of transactions and own them from design to production, in Java, Spring Boot, Kafka and AWS.

See my projects Contact me
Query my profilePick an endpoint
GET/v1/utkarsh/profile
{
  "name": "Utkarsh Kumar",
  "role": "Software Development Engineer 2",
  "company": "Nium",
  "experience_years": "2+",
  "focus": [
    "backend services",
    "distributed systems",
    "event-driven architecture",
    "microservices at scale",
    "cloud-native infrastructure"
  ]
}
200 OK18 ms0.3 kB

Highest-rated in my performance review, plus two CTO-level awards.

Nium, 2026
Performance review

High Impact

Nium's highest rating, for consistently delivering on or before deadlines.

CTO-level award

Star Performer

For delivering high-priority programs ahead of schedule.

CTO-level award

Rising Star

Awarded in the same year for delivering priority programs early.

1.3M+

transactions a month processed on the migrated SQS pipeline

1.5M+

daily transactions handled by the subscription engine

300+

engineers served by the AI incident investigation platform

98%

of crypto transactions cleared automatically through risk screening

Projects

Systems built and running in production.

Nium

AI incident investigation platform

When something breaks, this is where teams start. Built from scratch, it includes an MCP server with read-only tools over logs, databases, code and tickets for 27 backend services, a dashboard, and a Slack bot. Over 300 engineers and nearly 200 business users across 17+ teams use it.

80%faster, 30–45 min down to 5–10
MCPGrafana LokiPostgreSQLSlack API
Slack bot Dashboard AI + MCP Loki logs PostgreSQL Bitbucket Jira Confluence read-only tools
Nium

Transaction processing, moved from RabbitMQ to SQS

The queue that carries transactions into compliance was migrated with message ordering intact and Redis-backed deduplication added, so the same compliance decision is never made twice. It now handles peaks above 65K a day.

1.3M+transactions a month
AWS SQSRedisOrderingIdempotency
RabbitMQ SQS Redis dedupe Compliance duplicate dropped ordered, once only
MobiKwik

Subscription engine on the payout system

Designed a subscription engine wired into MobiKwik's payout system, serving 95K+ merchants. Platform and subscription fees were added, lifting revenue, along with an auto-retry job that settles failed payouts on its own and cut failures by 30%.

1.5M+transactions a day, 95K+ merchants
JavaSpring BootMySQLCron
Subscriptions Payout 95K+ merchants Auto-retry failures down 30% fees added
Nium

Blockchain risk screening

A service on AWS SQS that checks wallets and transactions in parallel against configurable risk rules. It has screened more than 90K crypto transactions, and nearly all of them clear without a human looking.

98%cleared automatically, 90K+ screened
AWS SQSParallel checksRules engine
Tx Wallet check Tx check Rules 98% auto 2% review in parallel
Nium

Singapore's government digital identity on FAPI 2.0

Integrated the national identity system under the FAPI 2.0 security profile, with eight protection layers between our app and the government. They include PKCE, DPoP token binding, encrypted ID tokens and replay prevention.

8protection layers on every request
FAPI 2.0OAuth 2.0PKCEDPoP
App Gov ID PKCE Encrypted ID DPoP Replay guard +4 more layers
Nium

Business verification service, built from scratch

A new Spring Boot microservice that verifies businesses at onboarding. It authenticates with outside vendors, takes their results back through asynchronous webhooks, stores documents, and caches responses so repeat checks stay fast.

0 → 1designed and shipped end to end
Spring BootWebhooksAWS S3Caching
Business Verify service Vendor async webhook Docs S3 Cache
Experience

Four teams, one thread: backend systems that handle money and scale.

  • Migrated transaction processing from RabbitMQ to SQS: 1.3M+ a month, peaks of 65K+ a day, with ordering and Redis-backed deduplication.
  • Integrated Singapore's government digital identity system on FAPI 2.0 with 8 protection layers.
  • Designed a new Spring Boot microservice for business verification at onboarding: vendor auth, async webhooks, document storage and response caching.
  • Built a blockchain risk screening service that has screened 90K+ crypto transactions and clears 98% automatically.
  • Created an AI incident investigation platform for 300+ engineers across 17+ teams, cutting diagnosis time by 80%.
  • Modernized 5+ production microservices from Java 8 to Java 21, with Spring Boot upgrades and Snyk fixes.
  • On-call for onboarding, compliance, document processing and risk systems, fixing concurrency, duplicate-processing and migration failures at the root.
  • Reduced API latency with Guava-based caching, cutting load from redundant sandbox API calls.
  • Designed a Lifecycle Policy Framework for flexible card behaviour: activation, reissuance and replacement.
  • Centralized product parsing and configuration so new card products onboard faster across regions.
  • Refactored card validation into a predicate-based model for tenant validations.
  • On-call engineer for credit and debit card issuance during Global Fintech Fest, owning stability and incident response at peak.
  • Built observability dashboards on Prometheus and Kafka Connect for delivery success, lag and health.
  • Designed a scalable Subscription Engine integrated with the Payout System, serving 95K+ merchants and 1.5M+ transactions daily.
  • Implemented platform and subscription fees, contributing to a significant revenue increase.
  • Built an auto-retry cron in Spring Boot that automated settlements and cut failed transactions by 30%.
  • Migrated merchant ID storage from comma-separated values in PostgreSQL to a normalized MySQL schema, reducing payout errors.
  • Kept payouts available and low-latency at peaks of millions of transactions a day.
  • Added personalized and static discount coupons to email campaigns, lifting conversion by 20.68%.
  • Migrated inbound SMS conversation handling from Ruby to Java.
  • Helped build automatic agent assignment across WhatsApp, SMS and Instagram.
Skills

Tech stack, grouped by layer.

Highlighted in red: the tools used most, every day, in production.

Languages
JavaPython
Backend and messaging
Spring BootKafkaMicroservicesSpring CloudApache CamelRabbitMQREST APIs
Cloud and DevOps
AWS SQSAWS S3AWS EKSAWS IAMDockerKubernetesJenkinsArgo CD
Databases
PostgreSQLRedisMySQLMongoDBLiquibase
Security
OAuth 2.0FAPI 2.0JWT
Observability
PrometheusGrafanaLokiKibana
AI tooling
MCP
Education
M.Tech, Information Technology
Netaji Subhas University of Technology
Aug 2022 – Jun 2024
B.Tech, Computer Science and Engineering
Asansol Engineering College
Aug 2017 – May 2021
Contact

Please contact me at.

Email utkarsh382@gmail.com
Phone +91 93041 14970