manojOS · boot sequence
0%
welcome to manoj's desk _
draw mode · press ESC to stop
🐛 catch the bugs!
20s 🐛 caught: 0 missed: 0
quack 🦆
~ ~ ~
DEPLOY
FRIDAY
NIGHTS.
(kidding. please don't.)
,MK '26
hey recruiter 👋
scroll? no. click
the laptop.
TODO:
▢ write tests
☑ break things
☑ fix things
remember: every
webhook wants to
retry. let it.
HOURS UNTIL
THE WEEKEND
·
DDIA
ch. 8: the trouble with distributed systems
...clocks lie. networks lie. so does your monitoring...
9:41
slack
kafka is fine
$ hello, friend
click to enter

✨ secret drawer

🐛 play bug-hunt
🎉 throw confetti
🎨 party mode
🔄 reset the desk
📝 leave a note

Manoj's Desk

👆 click the laptop for the "real" portfolio
crafted by manoj
manoj kumar
designed and engineered by hand · 2026
brewed on too much coffee · powered by curiosity
manoj006kumarr@gmail.com
coffees347
uptime99.9%
secrets0/38
resume.pdf
MANOJ KUMAR

Summary

Software engineer with several years of experience building backend services and distributed systems that have to stay up under real load. I work mostly in Java and Spring Boot on the server side, with Node.js, TypeScript, and React when the work reaches the front end, and I lean on Kafka, AWS, and a mix of SQL and NoSQL stores to keep things fast and reliable. Lately a good part of my work has involved wiring LLMs into production systems (recommendations, support automation, RAG) in ways that actually hold up rather than just demo well. I like the hard parts: event-driven design, fault tolerance, and getting latency and cost down without breaking things.

Skills

  • Languages: Java, JavaScript, TypeScript, C#, C++, Python, SQL
  • Backend: Spring Boot, Spring MVC, Microservices, Node.js, Express, Hibernate, REST & GraphQL APIs, event-driven architecture
  • Frontend: React.js, Redux, HTML5, CSS3
  • Cloud & Infra: AWS (Lambda, SQS, SNS, DynamoDB, CloudWatch, API Gateway, EC2, SES, IAM, CloudFormation), Docker, Kubernetes, Azure Key Vault, serverless
  • Data & Caching: DynamoDB, MongoDB, PostgreSQL, MySQL, Redis, Ehcache, sharding, replication
  • Messaging: Kafka, Amazon SQS/SNS, RabbitMQ
  • DevOps & Tooling: Jenkins, GitHub Actions, CI/CD, Git, ELK, Prometheus, Grafana, JUnit, Mockito, Swagger/OpenAPI, Smithy
  • AI & Automation: LLMs, RAG, embeddings & vector search, prompt engineering, tool calling, agent architecture, Selenium, Playwright, webhooks
  • Practices: distributed systems, system design, performance optimization, fault tolerance, TDD/BDD, design patterns, Agile
  • Security: OAuth 2.0, RBAC, JWT, encryption, secure API development

Experience

Software Development Engineer · Amazon | Arlington, TX Jul 2025 - Present
  • Build subscription lifecycle services in Java, Spring Boot, and AWS serverless that handle cancellations, suspensions, and partner agreements for the Amazon Fuse ecosystem, the plumbing behind Prime, Music, and Video offers across multiple regions.
  • Wrote the event-driven reconciliation platform on Lambda, S3, DynamoDB, and Glue that processes 10K+ financial transactions a day, automating billing validation and cutting the related infrastructure cost by about 90%.
  • Designed fault-tolerant microservices across Java, Node.js, and TypeScript using Kafka, retries, and circuit breakers so high-volume subscription events keep flowing even when something downstream is having a bad day.
  • Stood up the BFF layer in Node.js and GraphQL that backs the React + TypeScript apps, with feature flags and regional compliance rules baked in so product teams can ship without re-plumbing the backend.
  • Integrated LLM-backed recommendation and support services (Python, AWS Bedrock / OpenAI, RAG) for subscription insights and automated partner query resolution, useful enough to take real work off the operations team.
  • Added distributed caching on Redis and DynamoDB for entitlement and subscription lookups, which dropped API latency by roughly 40% on customer-facing paths.
  • Mentor a couple of newer engineers through design reviews and code walkthroughs, and migrated legacy orchestration services into containerized microservices on Docker and Kubernetes along the way.
Software Engineer · Adobe | Bangalore, India Mar 2021 - Jul 2023
  • Built and grew backend services in Java, Spring Boot, MySQL, and Kafka behind Adobe Experience Cloud and enterprise content workflows, where reliability and response time were the whole point.
  • Moved from feature-level work to owning backend modules end to end: APIs, integrations, and reusable services other teams could build on.
  • Designed distributed event-driven services with Kafka and Redis so content workflows, analytics pipelines, and customer activity stayed in sync across enterprise apps.
  • Wrote REST APIs and integrations across Node.js, Express, and Spring Cloud to make Adobe products and third-party systems talk to each other cleanly.
  • Built Python utilities for reporting and automation that gave the team better visibility into how the platform was actually being used.
  • Took part in the system-design discussions around microservice decomposition and resiliency, and tightened up deployments with Docker, Kubernetes, Jenkins, and GitHub Actions.
Software Engineer · Capital One | Remote, India Jan 2019 - Feb 2021
  • Built backend services in Java, Spring Boot, MySQL, and PostgreSQL for credit-card servicing, account management, and transaction processing, my first real exposure to financial systems at scale.
  • Worked on Kafka-based event-driven systems for transaction monitoring and payment workflows, which is where distributed systems stopped being theory for me.
  • Built reusable frontend modules in React, Redux, JavaScript, HTML5, and CSS3 for internal servicing dashboards and customer-facing banking apps.
  • Tuned backend performance with SQL query optimization and Redis caching to keep customer-facing financial workflows responsive.
  • Handled secure API work with Spring Security, OAuth2, and JWT, and shipped through Docker, Jenkins, AWS EC2, and CI/CD.
  • Wrote Python automation for reporting and log analysis, and spent real time on debugging and incident resolution, the kind of work that teaches you how systems actually fail.

Education

Master of Science · University of Texas at Arlington

Notes File Edit View Help
🔋
,
~/manoj.dev/notes.txt
~ notes ~

Hi, I'm Manoj.

welcome to my desk ↓

Manoj Kumar
me, probably.
status
open to roles
experience
5+ years
Software engineer who likes the hard parts: distributed systems, subscription & payments infra, event-driven systems. Several years shipping backend services at scale, currently building subscription infrastructure at Amazon.

I'm a software engineer who builds backend services and distributed systems that have to hold up under real load. Most of my work is on the server side (Java, Spring Boot, Node.js, TypeScript, Kafka, REST/GraphQL) but I'll go full-stack with React when the work reaches the front end. Cloud-native by default: AWS, Docker, Kubernetes, CI/CD.

I've done this at Amazon (subscription lifecycle and reconciliation infrastructure for the Fuse partner ecosystem behind Prime, Music, and Video), Adobe (backend services for Experience Cloud and enterprise content workflows), and Capital One (credit-card servicing and transaction processing). Different domains, same underlying problems.

Things I care about: distributed, event-driven systems, AI woven into the backend (RAG, tool calling, LLM integration that holds up in production), observability that catches incidents before they wake anyone up, and getting latency and cost down without breaking things. I'm the engineer who cut API latency 40% and reconciliation infra cost 90% by actually measuring where the work was going.

currently
SDE at Amazon. Building subscription lifecycle services, an event-driven reconciliation platform, and LLM-backed support tooling. Open to conversations about hard distributed-systems and AI infra problems.
~ m.

by the numbers

0
financial txns reconciled daily at Amazon
0
reconciliation infra cost cut on AWS serverless
0
API latency cut with Redis + DynamoDB caching
0
regions served by subscription services
manoj@desk:~/playgrounds$ ./run --interactive
loading 3 live demos...

click things. break things.

three systems problems you'll hit the moment you have more than one user. payments, healthcare, messaging, analytics , same hard parts. try to break them.

01. webhook idempotency

LIVE

any webhook source , Stripe, GitHub, Twilio, Shopify, your own services , will retry if you don't ack in time. your job: process exactly once. strategy: every event carries an event_id, you cache what you've processed. try hitting "duplicate" a bunch.

0processed
0dropped
$0total charged

02. exponential backoff + jitter

LIVE

some downstream service , a database, a payment processor, an EHR, a third-party API , flakes out 60% of the time. retry up to 5 times with base · 2^n + jitter. jitter prevents the thundering herd. every run is different.

0attempts
0mselapsed
,result
// hit run to start

03. token bucket rate limiter

LIVE

every API needs one. capacity 10, refill 2/sec. each request spends a token. empty → 429. try to spam faster than it refills.

0200 ok
0429
10tokens
bucket10/10
manoj.dev work experience

where i've worked.

3 roles · 5+ years · subscriptions, financial systems, and platform work.

Software Development Engineer
Amazon
'25 → now
10K+
txns/day reconciled
-90%
infra cost
-40%
API latency

When I joined Amazon, the brief was deceptively simple: keep subscriptions for the Fuse partner ecosystem (the machinery behind Prime, Music, and Video) running without a hitch across regions. The reality was messier. I built the lifecycle services on Java, Spring Boot, and AWS serverless, then wrote the event-driven reconciliation platform on Lambda, S3, DynamoDB, and Glue that quietly checks the books on more than ten thousand transactions a day. Along the way I leaned on Bedrock and RAG to take grunt work off the support team, and squeezed customer-facing latency down with Redis and DynamoDB caching. The work I'm proudest of is the stuff nobody notices, because it just works.

Software Engineer
Adobe
'21 - '23
2yr
to module owner
99.9%
service uptime
5+
teams unblocked

Adobe is where I grew from "give me a ticket" to "give me the module." I started writing backend services on Java, Spring Boot, MySQL, and Kafka behind Experience Cloud, and over two years ended up owning whole pieces of it end to end. I designed the event-driven flows on Kafka and Redis that kept content, analytics, and customer activity in sync across enterprise apps, wrote the REST APIs that tied Adobe products to the outside world, and made deployments boring (in the good way) with Docker, Kubernetes, and GitHub Actions. It taught me that owning something means caring about it after it ships.

Software Engineer
Capital One
'19 - '21
0
reconciliation defects
-30%
query latency
100%
PCI-compliant APIs

Capital One was my first real taste of money moving through code, and it made me serious about getting things right. I built backend services on Java, Spring Boot, MySQL, and PostgreSQL for credit-card servicing, account management, and transaction processing, where a rounding error isn't a bug, it's someone's bank statement. I worked on the Kafka event-driven systems watching transactions, tuned the slow paths with SQL and Redis, locked the APIs down with Spring Security, OAuth2, and JWT, and even built the React and Redux dashboards the servicing teams lived in. This is where distributed systems stopped being a textbook chapter for me.

things I can actually do.

grouped by what I reach for. not a wishlist. not a buzzword cloud.

🧠 languages
Java Python C / C++ JavaScript TypeScript SQL
🎨 frontend
React Redux Next.js RxJS Tailwind HTML/CSS
⚙️ backend
Spring Boot Spring Cloud Node.js Express REST GraphQL
🤖 AI / ML
LLMs RAG AWS Bedrock OpenAI APIs Tool calling Embeddings & vector search Agent architecture Prompt engineering
📡 data & messaging
Kafka Redis PostgreSQL MySQL MongoDB
☁️ cloud & devops
AWS Docker Kubernetes GitHub Actions CI/CD
🎯 production practices my thing
Idempotency Rate limiting Retries with jitter Circuit breakers OAuth2 / JWT Observability

things i've built.

7 items
● personal ● worked at
personal
🔍
reviewlens
a voice-of-customer decision engine.
turns customer reviews and feedback into actionable insights. built to surface what users actually think, prioritize the signal, and help teams make data-driven product decisions.
+ tap to learn more
personal
🪐
code-orbit
a playground for exploring code concepts visually.
an experimental space for visualizing algorithms, data structures, and computer-science ideas in the browser. designed to make abstract things tangible.
+ tap to learn more
personal
🎮
arcade-game
a browser arcade game, built just for fun.
a classic arcade-style game running entirely in the browser. demonstrates game-loop fundamentals, collision detection, and responsive input handling.
+ tap to learn more
worked at
📦
amazon
subscription & reconciliation infrastructure.
the subscription engine behind Prime, Music, and Video offers across the Amazon Fuse partner ecosystem, handling cancellations, suspensions, and partner agreements at scale. Built an event-driven reconciliation platform that quietly checks the books on 10,000+ transactions a day, cut infrastructure cost by ~90%, and trimmed customer-facing latency by 40%. Added LLM-backed support tooling that takes grunt work off the team.
+ tap to learn more
worked at
🎨
adobe
backend for experience cloud at enterprise scale.
backend services behind Adobe Experience Cloud and its enterprise content workflows, where reliability and response time were the whole point. Grew from feature work to owning entire modules end to end, designed the event-driven flows that keep content, analytics, and customer activity in sync across enterprise apps, and made deployments boring in the best way. Unblocked five-plus teams and held 99.9% uptime.
+ tap to learn more
worked at
🏦
capital one
credit-card servicing & transaction processing.
the systems behind credit-card servicing, account management, and transaction processing, where a rounding error isn't a bug, it's someone's bank statement. Built the event-driven monitoring that watches transactions in real time, tuned the slow paths until query latency dropped 30%, locked every API down to PCI-compliant standards, and shipped the dashboards the servicing teams lived in. Zero reconciliation defects.
+ tap to learn more
personal
this site
an interactive desk portfolio you're using right now.
an interactive desk experience with draggable items, a fake operating system inside the laptop, three live playgrounds, achievements, drawing mode, and a hidden drawer. you're inside it.
+ tap to learn more
🎧
NOW PLAYING
life outside work
manoj · hobbies playlist
0:00

manoj's life outside work

the playlist that runs when the laptop is closed · 7 tracks

1
🍜
cooking new recipes
the kitchen sessions
1,847 plays
4:32
2
🎮
video games
console and keyboard
2,398 plays
6:12
3
📺
tv, but responsibly
one more episode
1,204 plays
3:48
4
💪
the gym
mon/wed/fri, mostly
588 plays
3:02
5
🌐
building websites for fun
sites like this one
412 plays
5:17
6
🍽
exploring restaurants
a new place every week
306 plays
2:48
7
✈️
weekend travels
anywhere < 4 hours away
156 plays
5:21
Manoj Kumar
Manoj Kumar ● online · usually replies fast
hey manoj 👋
saw your portfolio. you made a desk? today · just now
yep. it was more fun than a landing page. today · just now
fair. we're hiring senior full-stack. strong systems engineers only. interested? today · just now
definitely. let's talk. today · just now
reply times are fast. usually.
🗒Notes
Terminal
📂Finder · Experience
🛠️Projects
🎧Music · Life
💬Messages