Hi, I am
Shivanand Vishwakarma

3.5+ yrs: production experience
1,000+: interviews / month served
40–60%: faster MongoDB queries
~55%: faster CI/CD builds

ABOUT ME

Hi there! I'm Shivanand Vishwakarma, a Full Stack Developer specializing in the MERN stack, with 3.5+ years of experience building scalable, high-performance SaaS products.

I specialize in building end-to-end web applications using React.js, Next.js, Redux Toolkit, TypeScript, Node.js, Express.js, MongoDB, and Redis. I enjoy creating intuitive, responsive user interfaces while designing robust backend systems that are scalable, secure, and production-ready.

Over the past 3.5+ years, I've contributed to AI-powered interview platforms, enterprise SaaS applications, and IoT-enabled products. From developing reusable React components and optimizing frontend performance to designing REST APIs, real-time systems, scalable backend architectures, and CI/CD pipelines for reliable deployments, I enjoy solving complex engineering challenges that create measurable business impact.

Currently, I'm also building production features in Python and PHP/Laravel alongside the MERN stack.

I'm passionate about clean code, performance optimization, system design, and continuous learning. Outside of work, I enjoy building personal projects, exploring new technologies, and following cricket.

Education: B.E. Mechanical Engineering (SPPU, 2017–2021) · Full-Stack Web Development, Masai School (2021–2022) — I switched from mechanical engineering to software and haven't looked back.

Bengaluru · Open to Full Stack / Backend roles

Let's talk My Resume GitHub LinkedIn

WORK HIGHLIGHTS

Production systems I've designed and shipped over 3.5+ years

★ Built end-to-end · solo

Interview Screen Recording & Offline-Safe Chunked Upload

AI interview platform · high-volume video recordings

Flow: browser records the interview in chunks → upload API → GCP/S3; offline chunks buffered in IndexedDB and flushed first on reconnect; grace-period merge into a seekable video, late chunks re-merged; final video used for review

Replaced a paid third-party recording serviceOwned frontend → backend → storageSurvives network drops (IndexedDB buffer)Always ends in a seekable full-length video

What I built

Designed and built an in-house replacement for a paid third-party recording service, removing the vendor dependency for interview recordings.
Frontend (React): records the live interview in timed chunks and uploads each one while the interview is still running, with live upload progress and graceful error states.
Offline-safe: when the network drops, new chunks are buffered in IndexedDB; on reconnect the backlog uploads first, then live chunks continue — nothing is lost.
Backend (Node.js/Express): chunk APIs store every part by index in GCP Cloud Storage / AWS S3 and track which chunks have arrived.
On interview end, the merge waits a grace period for missing chunks, then builds a seekable video; chunks that arrive late trigger a re-merge into a new full-length video.
The final recording powers review workflows — interview feedback, plagiarism detection and more.

ReactIndexedDBNode.jsExpressGCP Cloud StorageAWS S3MongoDB

Interview Lifecycle APIs & Query Optimization

Interview-tech platform · core backend

40–60% faster queries~25% faster feature delivery

What I built

REST APIs covering onboarding → evaluation → verdict, with role-based access control for candidates, interviewers, and admins.
Compound indexes and aggregation pipelines for heavy reporting queries; Redis caching for hot interview/session data.
Reusable middleware and React/Next.js/Redux component library shared across features.

Node.jsExpressMongoDBRedisNext.jsRedux Toolkit

Live Session Reliability & CI/CD

Real-time interviews · DevOps

~60% smaller Docker images~55% faster CI/CD builds

What I built

WebSocket retry and reconnection policies that keep live interview sessions alive through brief disconnects.
Contributed to Docker image and CI/CD pipeline optimizations for faster, lighter deployments.

WebSocketsSocket.IODockerCI/CD

AI-Driven Async Video Interview Platform

AI interview product · microservices

1,000+ interviews / month20–30% faster dashboards

What I built

Node.js/Express microservices for scheduling, dynamic question flows, and automated scoring logic with RBAC.
Refactored React/Redux dashboard modules for faster loads and smoother interactions.

Node.jsExpressMicroservicesReactNext.jsRedux

Hotel Management SaaS + Smart-Room IoT

Hospitality SaaS · built from scratch

100+ rooms in production4+ IoT device types

What I built

Booking engine, front desk, room management, and admin modules for a 100+ room property.
Event-driven IoT integration via Tuya APIs and webhooks — lighting, thermostat, presence sensors and more, with real-time device state sync.
Firebase-powered live dashboards, secure auth flows, code splitting and lazy loading.

ReactNode.jsMongoDBTuya APIsWebhooksFirebaseView live ↗

TECH SKILLS

Frontend

DevOps & Tools

Backend

Cloud & Deployment

Languages

Real-time & Integrations

Also Working With

PROJECTS

Collab Docs
PROJECT DETAILS
Description: A production-grade local-first collaborative document editor — combining offline editing, real-time sync, version control, secure multi-user access, and AI-assisted writing.
Features:
- Local-first architecture — edit documents offline with IndexedDB as primary storage, zero network blocking
- Background sync engine — automatic merge on reconnect without overwriting offline work
- Real-time collaboration over WebSockets with Y.js YATA deterministic conflict resolution
- Rich text editing with Tiptap 3 and real-time collaboration extensions
- Version history & time travel — save named snapshots and restore without corrupting live collaborators
- Invite collaborators by email with Owner, Editor, and Viewer role management
- Role-based access enforced at WebSocket level — viewers cannot push document updates
- AI writing assistant — continue writing, summarize, fix grammar via Groq Llama 3.3 70B
- Production safeguards — NextAuth JWT auth, Zod payload validation, 5 MB sync limits, and tenant-scoped MongoDB access
- Split production architecture — Next.js on Vercel + Y.js WebSocket server on Render, with GitHub Actions CI/CD
Tech used:
- Next.js 16
- TypeScript
- React
- Node.js
- MongoDB
- Mongoose
- Tailwind CSS
- Tiptap
- WebSockets / Y.js
- NextAuth
- Zod
- CI/CD
- Groq AI
SEE LIVE SOURCE CODE
Vybe Cabs
PROJECT DETAILS
Description: A ride-hailing backend that finds the nearest available drivers, notifies several at once over WebSockets, and guarantees exactly one of them is assigned, even when multiple drivers accept at the same instant.
Features:
- Nearby-driver search with Redis GEO (GEOSEARCH within a radius)
- Race-free acceptance: a single atomic Redis Lua script (SET NX) picks exactly one winner; the rest get 409
- BullMQ delayed jobs handle offer timeouts and retry with a wider search radius (3 km → 6 km → 10 km)
- Explicit ride state machine: REQUESTED → SEARCHING → ASSIGNED / TIMEOUT / CANCELLED, with a per-round audit trail
- Idempotent accept: retrying a winning request returns the same result
- Automated concurrency test script proving the exactly-one-winner guarantee
Tech used:
- NestJS
- TypeScript
- Redis
- BullMQ
- MongoDB
- Socket.IO
- Docker
SOURCE CODE
Serviots
PROJECT DETAILS
Description: A full-stack co-working space booking platform: members browse desks and meeting rooms and book time slots, while admins manage spaces and bookings. Double-booking is prevented at the database level.
Features:
- No double-booking: a unique MongoDB index on space + date + slot, with compensating rollback (verified with 8 concurrent requests → 1 success, 7 × 409)
- JWT access tokens kept in memory, plus httpOnly refresh-token rotation with reuse detection
- Role-based access for members and admins
- Rate limiting and request validation on the API
- Email notifications for bookings
- One-command Docker Compose setup; deployed on Vercel (frontend) + Render (API) + MongoDB Atlas
Tech used:
- React
- Node.js
- Express
- MongoDB
- Docker
- Vercel
Free hosting — first load ~30s
SEE LIVE SOURCE CODE
E-Commerce
PROJECT DETAILS
Description: A full-stack MERN commerce platform that handles 10k+ SKUs with sub-200ms API responses. It covers the full order lifecycle, from browsing to Stripe checkout.
Features:
- Compound MongoDB indexing keeps catalog queries under 200ms across 10k+ SKUs
- JWT-based authentication with protected routes
- Stripe payment integration and full order lifecycle
- Real-time inventory sync between cart and stock
- Product filters (size, color) and sorting
- Cart state managed with Redux Toolkit
Tech used:
- Node.js
- Express
- MongoDB
- React
- Redux Toolkit
- Material UI
- Styled Components
Free hosting — first load ~30s
SEE LIVE SOURCE CODE
Chat App
PROJECT DETAILS
Description: A real-time messaging platform built on Socket.IO that supports 100+ concurrent users across one-on-one and group conversations, with graceful reconnection when the network drops.
Features:
- Real-time one-on-one and group messaging over Socket.IO, built for 100+ concurrent users
- Graceful WebSocket reconnection: sessions recover automatically after network drops
- Live typing indicators and unread-message notifications
- Group admin controls: create, rename, add and remove members; members can leave
- Sign-in/sign-up authentication and user search by name or email
- Profile and group-info modals; image uploads via Cloudinary
Tech used:
- Node.js
- Express
- MongoDB
- React
- Cloudinary
- Socket.IO
- Chakra UI
Free hosting — first load ~30s
SEE LIVE SOURCE CODE