tgtushar@systemsbackend command center
/home/tushar/backend-systemsnode: production

backend engineer / distributed systems / operational tooling

Building backend systems, event-driven infrastructure, and operational tooling at scale.

Backend engineer with 3+ years of experience building scalable APIs, distributed systems, production backend services, logging platforms, and commerce infrastructure.

GitHubResumeLinkedInContact
100k+community members supporteddistributed Discord infrastructure
200k+daily events processedmessages, moderation actions, voice activity
+30%ONDC order volume liftAPI work at Shiprocket
~70%faster API responsesRedis-backed caching paths
API contractsevent ingestioncache designaudit logsbackground jobsoperator workflows

Systems I've built

Infrastructure, workflows, and backend surfaces I have operated.

This section is intentionally not a project gallery. It is a scan of the systems, scale, constraints, and production responsibilities behind the work.

Observability100k+ members, 200k+ daily events

Discord Logging Platform

Built distributed logging and moderation infrastructure that processes messages, moderation actions, and voice activity for real-time operational insight.

Large Discord communities
API Platform+30% order volume

ONDC API Layer

Led scalable REST API development for ONDC order flow, tracking, operational event monitoring, and logistics-side service communication.

Shiprocket
Performance~70% response-time reduction

Redis-backed API Paths

Implemented Redis-based caching for high-traffic API paths while keeping source-of-truth services and invalidation behavior explicit.

Shiprocket
Internal PlatformProduction commerce workflows

Commerce Workflow Automation

Designed delayed payment disbursement, automated auction reassignment, scheduled financial jobs, ERP integrations, and reporting workflows.

Magnolia Pearl
Trust and SafetyReal-time moderation insights

Moderation Infrastructure

Designed event ingestion and search workflows for real-time moderation insights, operational monitoring, and operator-facing automation.

Discord communities
Real-time SystemLive voice workflows

Private Voice Channel System

Created coin-based private voice channels with billing modes, access control, ownership transfer, and lifecycle cleanup.

Discord communities

System dossiers

Architecture, scale, decisions, and operational concerns.

Each dossier explains the backend problem, why the architecture mattered, and how the system behaves under real operational pressure.

system dossier / 01

Community Infrastructure and Logging Platform

Backend owner for Discord event ingestion, logging, and workflow design

100k+ members200k+ daily eventsreal-time moderation insights

Problem

Large Discord communities need message activity, moderation actions, and voice events to be ingested, searched, and explained fast enough for live operators.

Architecture

Discord gateway events flow through ingestion handlers, moderation workflows, Redis-backed operational paths, MongoDB persistence, and operator-facing insight surfaces.

Scalability / Reliability

Separated interaction handling from event ingestion, kept moderation workflows permission-aware, and shaped event records for search, debugging, and operational monitoring.

Operational Complexity

  • Built distributed logging and moderation infrastructure for communities with 100k+ members.
  • Processed 200k+ daily events across messages, moderation actions, and voice activity.
  • Made moderation activity easier to search, explain, monitor, and automate.

Technical decisions

  • Designed event ingestion around messages, moderation actions, and voice activity.
  • Stored searchable context for real-time moderation insights and later review.
  • Implemented backend automation systems and operational monitoring tools around event-driven workflows.
Node.jsMongoDBRedisDiscord.jsevent ingestionoperational monitoring

system dossier / 02

ONDC / Shiprocket Backend Systems

Graduate Engineer Trainee - backend APIs and integration workflows

+30% order volume~70% response-time reduction500+ daily live ordersS3-backed logs

Problem

Logistics order flow requires fast API responses, reliable tracking data, centralized logs, and integration paths that can handle marketplace and fulfillment-side constraints.

Architecture

ONDC-facing REST APIs connected order flow, Redis cache paths, structured S3 logs, monitoring workflows, and dashboard surfaces for operational visibility.

Scalability / Reliability

Improved read performance with Redis-backed paths and moved operational logs into durable storage for inspection without slowing request handling.

Operational Complexity

  • Improved ONDC order volume by 30%.
  • Reduced API response times by roughly 70% using Redis-based caching.
  • Built real-time monitoring workflows for 500+ daily live orders and operational events.

Technical decisions

  • Kept frequently accessed order and tracking reads behind Redis.
  • Centralized structured application logs in Amazon S3 for monitoring and debugging.
  • Built real-time monitoring workflows around live order and operational event activity.
Node.jsExpress.jsMongoDBRedisAmazon S3REST APIs

system dossier / 03

Magnolia Pearl Backend Workflows

Software Developer for Magnolia Pearl commerce backend workflows

PayPal delayed disbursementscheduled background jobsERP integrations

Problem

Commerce operations needed backend workflows for delayed payment disbursement, auction reassignment, financial processing, reporting, and external ERP integrations.

Architecture

ASP.NET and Razor Pages workflows coordinate PayPal delayed disbursements, scheduled background jobs, SQL reporting paths, Azure Blob Storage, and third-party ERP APIs.

Scalability / Reliability

Designed transactional workflow boundaries for bidding and financial operations so scheduled jobs, reporting, and external integrations can evolve without fragile handoffs.

Operational Complexity

  • Owned production backend workflows for a commerce platform.
  • Automated financial processing, reporting, delayed disbursement, and auction reassignment workflows.
  • Integrated third-party ERP systems and external APIs into secure commerce infrastructure.

Technical decisions

  • Designed PayPal delayed-disbursement flows around explicit payment states.
  • Built automated auction reassignment logic with transactional consistency during bidding workflows.
  • Optimized SQL queries for business-critical financial reporting and operational review.
ASP.NETC#Razor PagesSQLAzure Blob StoragePayPal

system dossier / 04

Real-time Voice and Support Systems

Backend developer for real-time community systems

live PVC workflowscoin-based billingpermission-aware controls

Problem

Live communities need voice-channel lifecycle automation, access control, user support, and moderation feedback loops that work while people are actively using them.

Architecture

Gateway events drive voice-channel state, coin billing, membership permissions, support actions, and cleanup jobs.

Scalability / Reliability

Designed workflows around explicit ownership, membership state, and cleanup paths so real-time resources do not drift out of sync with user actions.

Operational Complexity

  • Enabled paid and private voice-channel workflows for Discord communities.
  • Reduced manual voice-channel management for moderators and owners.
  • Created clearer user support paths through panels, transcripts, and action history.

Technical decisions

  • Modeled private voice channels as resources with owners, members, billing mode, and lifecycle state.
  • Used permission-aware commands for add, remove, hide, transfer, public, private, and delete actions.
  • Kept support workflows close to moderation logs so operators can resolve issues quickly.
Discord gateway eventsNode.jsMongoDBaccess controlreal-time state

Architecture visualization

Request lifecycles, logging pipelines, and workflow maps.

These diagrams show the pattern language I use: validate the boundary, isolate the workflow, persist enough context, and make the system observable.

Event Ingestion Pipeline

A request path that keeps user-facing actions fast while preserving enough context for retries, audits, and operator review.

  1. 01Event Sourcewebhook, gateway, API request
  2. 02Validateschema, auth, idempotency key
  3. 03Dispatchworkflow handler or queue
  4. 04Persiststate change and audit log
  5. 05Observedashboard, transcript, alert
Logging and Audit System

Operational logs are designed as a product surface: queryable, explainable, and tied to the user or workflow that created them.

  1. 01Actionmoderation or order event
  2. 02Context Builderactor, target, correlation id
  3. 03Durable LogS3, database, transcript
  4. 04Review Surfacedashboard or support thread
Caching Layer

Cache hot reads close to the API while keeping source-of-truth ownership and invalidation rules explicit.

  1. 01Client Requestorder, tracking, profile
  2. 02API Contractbounded response shape
  3. 03Redisfast path for repeat reads
  4. 04Primary Serviceauthoritative state
  5. 05RefreshTTL or write-through update

Engineering notes

Backend lessons from operating systems, not just building them.

Short writeups that make the portfolio feel like a working notebook: cache semantics, logging quality, background jobs, and real-time state.

Latency and correctness

Redis Caching Is a Contract, Not a Shortcut

The cache has to make ownership explicit: what is authoritative, how entries expire, which paths can tolerate stale reads, and what operators should inspect when behavior diverges.

Observability

Event Logs Should Explain the System

A useful log stream carries actor, target, workflow, correlation id, and state transition. That makes moderation and order systems debuggable without reconstructing history from raw noise.

Workflow reliability

Background Jobs Need Product Semantics

Payment disbursement, auction reassignment, reporting, and cleanup jobs should map to clear business states so retries are safe and failures are visible to humans.

State reconciliation

Real-time Systems Fail Through Drift

Voice channels, permissions, and live moderation flows need cleanup and reconciliation paths. The hard part is keeping resources aligned after disconnects, retries, and manual intervention.

Technical inventory

The stack behind the systems.

Grouped by how backend work actually gets done: language, service layer, persistence, cloud primitives, and event-driven workflows.

Languages

JavaScriptTypeScriptC#JavaPythonSQL

Backend & Frameworks

Node.jsExpress.jsASP.NETRazor PagesSpring BootREST APIsMicroservices

Databases & Caching

MongoDBPostgreSQLMySQLRedisschema designquery optimization

Cloud & Infrastructure

AWSAzureGCPAmazon S3Azure Blob StorageGit

Real-time Systems

Discord gateway eventsSocket.IOvoice workflowsModMaillive monitoring

Event-driven Systems

event ingestionbackground jobswebhooksERP integrationsautomation workflows

Production timeline

Commerce, logistics, and community infrastructure.

Roles where I shipped APIs, scheduled jobs, integrations, logging paths, monitoring workflows, and operator-facing automation.

Jan 2024 - Present

Software Developer Magnolia Pearl

Texas, US / ASP.NET, C#, Razor Pages, SQL, Azure Blob Storage, PayPal

  • Designed and implemented backend workflows for delayed payment disbursement using PayPal APIs.
  • Built automated auction reassignment logic ensuring transactional consistency during bidding workflows.
  • Developed scheduled background jobs, automated financial processing workflows, reporting paths, optimized SQL queries, and secure ERP integrations.
Jun 2022 - May 2023

Graduate Engineer Trainee Shiprocket

New Delhi, IN / Node.js, Express.js, MongoDB, Redis, Amazon S3

  • Led development of scalable REST APIs for the ONDC initiative, contributing to a 30% increase in order volume.
  • Implemented Redis-based caching, reducing API response times by roughly 70%.
  • Built secure backend data pipelines, centralized structured application logs in Amazon S3, and real-time monitoring workflows for 500+ daily live orders.
Jul 2021 - Oct 2021

Software Engineer Trainee Roxiler Systems

Remote, IN / Node.js, Express.js, PostgreSQL

  • Assisted in developing and maintaining Node.js and PostgreSQL backend services.
  • Integrated Slack, Telegram, and Discord APIs for automated notifications and workflows.
Nov 2020 - Feb 2021

Backend Developer Intern Gocal

Remote, IN / Node.js, Express.js, MongoDB, Bcrypt, JWT

  • Designed backend architecture and REST APIs for a delivery platform supporting real-time inventory and scheduled deliveries.
  • Optimized database schemas to improve query performance.
  • Implemented JWT-based authentication, role-based authorization, scheduled deliveries, and real-time inventory management.

Operating philosophy

The rules I come back to when systems get real.

The more operational pressure a backend takes, the more it rewards clear ownership, explicit state, useful logs, and boring reliability.

Scalability

I separate hot paths from operational workflows, cache the right reads, and design APIs around bounded payloads, clear ownership, and predictable failure modes.

Reliability

Production code should make failure visible. I prefer explicit state transitions, durable logs, defensive integrations, and workflows that can be retried safely.

Maintainability

I optimize for future operators: small service boundaries, readable handlers, intentional schemas, and documentation that explains tradeoffs instead of restating code.

Backend Ownership

Owning a backend means caring about deploys, data integrity, observability, support workflows, and the humans who need to debug the system at 2 AM.

Observability

I treat logs, audit trails, dashboards, and transcripts as product surfaces. They shorten incidents, explain system behavior, and build trust with internal teams.

Credentials

Education and applied engineering wins.

A compact view of the academic and hackathon credentials behind the backend work.

Education

DIT University

Bachelor of Technology, Computer Science and Engineering

Dehradun, INAug 2018 - May 2022CGPA: 8.82/10

Award

Winner: Open Innovation Track

LNMHacks 5.0

Jan 2023

Award

Winner: NABARD-ONDC Hackathon Integration Challenge

ONDC ecosystem challenge

Jul 2022

Open a channel

Backend, platform, infrastructure, and production engineering conversations.

Focused on backend, platform, infrastructure, and production engineering roles. I am strongest where APIs, integrations, production debugging, and infrastructure ownership meet.

EmailGitHubLinkedInResume