Hi, I'm
Akash Gupta_

Artistic

Full Stack

Download CV

Developer

Download CV

Please, do NOT hold the button below!

About Me

What am I good at

⚠️ System Warning: Long-pressing items reveals classified information

Languages I use to create bugs faster

TypeScript, JavaScript, Python, C++, C

Making things look pretty while the backend burns

Next.js, React, React Native, Tailwind CSS, Zustand, Redux, Tanstack React Query, Framer Motion, GSAP

Where I do the heavy lifting that nobody ever sees

Node.js, Express.js, Flask, PostgreSQL, MongoDB, Prisma ORM, REST & GraphQL APIs

Tools I use to accidentally break production

Docker, AWS (S3), CI/CD, GitHub Actions, Linux, Git, Pandas, System Design

Yes, I can probably figure that out

Java, Data Science, NumPy, Basic ML

Experience

I've spent the last 2 years deciphering the legacy spaghetti code of Fortune 500 clients, backed by 5 years of aggressively googling stack traces, and console.logs

Node* head =

Aug 2024 — Present

Next.jsTypeScriptTailwind CSSSASSTanstack React QueryReduxAWS S3Java

—Migrated legacy AEM enterprise apps to Next.js (SSR & Server Actions), slashing page load times by over 90%.
—Achieved a 40% increase in overall delivery speed and a 95% reduction in translation time by developing and integrating custom agentic AI workflows, demonstrating real AI ROI.
—Built a highly scalable Next.js/AWS S3 data pipeline utilizing multipart chunked uploads to handle massive 50GB file transfers without timeouts.
—Developed a custom async state management library with timed caching, reducing API polling and optimizing the existing Redux architecture.
—Deployed secure Backend-for-Frontend (BFF) proxy APIs to interface with Java microservices, Apigee, and Salesforce.
—Architected a custom i18n routing system from scratch using advanced Next.js rewrites to bypass strict enterprise deployment constraints.
—Acted as the primary Next.js SME, mentoring 10+ engineers (including senior engineers) on modern React paradigms and resolving architectural blockers.

Jul 2023 — Dec 2023

React (Vite)PostgreSQLExpress.js

—Engineered an end-to-end web app using React, Express, and PostgreSQL, driving the full lifecycle from schema design to deployment.
—Implemented client-side performance optimizations including component lazy loading, dynamic routing, and optimistic state updates.

NULL

Wait, it's already reversed in chronological order!
O(1) complexity 😌.

OOPS! I forgot to remove the console log again 😫

Yes, go ahead and click it to inspect!

Projects

Built with high coffee-to-code efficiency, over-engineered state management, and an unhealthy obsession with making it look perfect and strictly typing everything so I never have to see any again.

01.

Hangout Lobby

Real-Time Chat & Community Platform

ZeroMessage Drop Rate

WebRTCSub-100ms Voice & Video Latency

InfiniteCursor-based Infinite Scroll

WebSocketsWebRTCSocket.ioTanStack QueryNext.js

A Discord-inspired communication platform engineered for low-latency collaboration. It features real-time text messaging, WebRTC-powered voice and video rooms, and cursor-based pagination to handle high-volume chat histories efficiently. Built with an automated fallback mechanism to ensure zero message drops during network disruptions.

Case Study

Expand

Problem Statement

Building a low-latency, highly reliable communication platform that handles complex state across text, audio, and video without crashing during connection drops.

01.Bypassed Next.js serverless limitations by custom-configuring a NextApiRequest hack to handle persistent bi-directional WebSockets.
02.Managed complex global state for audio/video permissions and active UI interactions using Zustand and custom React hooks (useChatSocket, useChatScroll).
03.Ensured secure and seamless media sharing utilizing UploadThing for robust file handling.

Project Links & Artifacts

Explore live production builds or view codebases on GitHub.

View Source Code

02.

AI Companion

Conversational AI SaaS Platform

13BModel Parameters

100%Session Memory Retention

ServerlessStripe Webhook Processing

Llama-2 13bLangChainPinecone Vector DBUpstash RedisReplicate

A comprehensive SaaS platform where users can configure and interact with highly customized AI companions. It leverages Llama-2 models with Pinecone vector embeddings to maintain long-term conversational memory. Includes a fully integrated Stripe subscription tier and a highly responsive, real-time chat interface.

Case Study

Expand

Problem Statement

Providing a scalable, monetized platform that enables non-technical users to easily engineer and interact with custom AI models while maintaining persistent conversational context.

01.Successfully managed LLM context windows by chunking and retrieving conversational history from the vector database before generating responses.
02.Ensured seamless, low-latency UI updates during model inference using custom hooks and streaming responses.
03.Designed a flexible Prisma schema to associate custom companions, chat histories, and Stripe subscription statuses accurately with Clerk user IDs.

Project Links & Artifacts

Explore live production builds or view codebases on GitHub.

View Source Code

03.

Multi-Tenant E-Comm CMS

Headless e-commerce CMS & Decoupled Storefront

IsolatedMulti-Store Data Contexts

DynamicRuntime REST Endpoints

HeadlessDecoupled Client Architecture

Multi-Tenant ArchitectureStripe PaymentsHeadless CMSMySQL (PlanetScale)TanStack Table

A decoupled e-commerce ecosystem featuring a central multi-tenant admin CMS and a standalone storefront. Generates dynamic isolated REST endpoints for distinct stores, backed by Stripe webhooks and real-time analytics dashboard. Enables third-party frontends to consume catalog APIs securely with zero database exposure.

Case Study

Expand

Problem Statement

Building a scalable, multi-tenant architecture where a single admin application can securely control data, media, and API endpoints for infinitely many independent storefronts.

01.Architected the initial schema in MySQL via PlanetScale, later executing a full database migration to MongoDB, adapting Prisma relational models to document structures to optimize infrastructure costs.
02.Designed complex relational data models in Prisma to ensure strict data isolation and cascade-deletion integrity between different stores.
03.Managed sophisticated client-side state across two separate repositories using Zustand for shopping carts and admin UI toggles.
04.Implemented an advanced data table system using TanStack Table to handle large inventory datasets with complex filtering and pagination.

Project Links & Artifacts

Explore live production builds or view codebases on GitHub.

Admin Dashboard Repo Storefront Repo

04.

Learnify

Comprehensive Learning Management System (LMS)

HLSAdaptive Bitrate Streaming

RBACRole-Based Access Control

Dual-StoreSanity + Prisma Architecture

Mux (HLS Video)RazorpaySanity CMSNext-Auth v5Uploadthing

An enterprise-grade learning management system designed for video-heavy curricula. It features role-based access control, drag-and-drop course builders, and granular student progress tracking. Powered by Mux for adaptive HLS video streaming and Razorpay for seamless, secure checkouts.

Case Study

Expand

Problem Statement

Engineering a scalable educational platform that handles heavy media streaming, complex access control logic, and a multi-tiered affiliate commission system.

01.Managed complex database relations to accurately calculate and persist granular student progress across hundreds of individual video lessons.
02.Handled secure video file uploads and optimized playback delivery through UploadThing while maintaining seamless YouTube embed fallbacks.
03.Upgraded and configured Next-Auth v5 (Auth.js) to manage complex session states across various user roles (Admin, Creator, Student, Affiliate).

Project Links & Artifacts

Explore live production builds or view codebases on GitHub.

View Source Code

Ready to deploy the featured projects?

Spin up a live production instance to verify the architecture.

Side Quests That Actually Survived

Unlike the other fifteen project ideas currently sitting abandoned in my Vercel dashboard.

Doc Chat

Interactive RAG Document Assistant

tRPCLangChainPinecone Vector DBRAG PipelineReact PDF

View Source Code

An intelligent document analysis tool that allows users to upload PDFs and interact with them through context-aware queries. Built on a Retrieval-Augmented Generation (RAG) pipeline using LangChain and Pinecone, with end-to-end type safety enforced by tRPC.

Stay Scape

Full-Stack Airbnb Clone

Reservation EngineMapbox APINext.jsNextAuthPrisma

Live Demo View Source Code

A full-stack travel booking application replicating Airbnb's core user flows. Supports multi-criteria search filtering, interactive property listings, favorite curation, and date-locked reservation management, alongside a seamless checkout experience and mobile-optimized UI.

Metaverses

Fluid Web3 Interactive Experience

Framer MotionSpatial UI DesignUI/UXNext.jsTailwind CSS

Live Demo View Source Code

A futuristic, high-performance UI showcase focused on modern web design. Built to demonstrate fluid micro-interactions, responsive cross-device layouts, and scroll-triggered animations powered by Framer Motion.

React Native Job Portal

Cross-Platform Mobile Job Board

React NativeExpoREST APIsZustand

View Source Code

A clean React Native mobile application designed to browse and filter jobs across global markets. Integrates external job APIs with dynamic search queries, salary filtering, and device-responsive experience across iOS and Android.

MyPass Vault

Desktop Password Manager

PythonTkinterSQLiteLocal Cryptography

View Source Code

A modular desktop password manager built with Python and Tkinter. Implements isolated window controllers, custom GUI theming via proxttk, and secure local data handling for user authentication and recovery flows.

//

// vim_session: try to quit

~".git/MERGE_MSG" 6 lines, 342 bytes

~quick minor fix (final_v2_really_final)

~# Please enter a commit message explaining why this merge is necessary.

~# Lines starting with '#' will be ignored, but your tech lead won't ignore this.

~# 47 files changed, 3,841 insertions(+), 2,490 deletions(-)

-- NORMAL --

Let's Connect

Just please don't ask me to fix your printer. I only debug servers and write code, provided there is a steady supply of caffeine ;)

Socials

GitHubGitHubGitHub

LinkedInLinkedInLinkedIn

InstagramInstagramInstagram

Location

Bengaluru, IndiaBengaluru, IndiaBengaluru, India

Available for full-stack opportunities in Bengaluru & remote globally.

Download CV

Email

akashgupta4510@gmail.comakashgupta4510@gmail.comakashgupta4510@gmail.com

Phone

+91 86177 31097+91 86177 31097+91 86177 31097

Hi, I'm
Akash Gupta_

Creative

Full Stack

Download CV

Developer

Download CV

Please, do NOT hold the button below!

About Me

What am I good at

⚠️ System Warning: Long-pressing items reveals classified information

Languages

TypeScript, JavaScript, Python, C++, C

Frontend & Mobile

Next.js, React, React Native, Tailwind CSS, Zustand, Redux, Tanstack React Query, Framer Motion, GSAP

Backend & Databases

Node.js, Express.js, Flask, PostgreSQL, MongoDB, Prisma ORM, REST & GraphQL APIs

Cloud, DevOps & Tools

Docker, AWS (S3), CI/CD, GitHub Actions, Linux, Git, Pandas, System Design

Familiar With

Java, Data Science, NumPy, Basic ML

Experience

I bring over 2 years of professional experience navigating the complex architecture of Fortune 500 clients, backed by 5 years of dedicated full-stack engineering.

Node* head =

Aug 2024 — Present

Next.jsTypeScriptTailwind CSSSASSTanstack React QueryReduxAWS S3Java

—Migrated legacy AEM enterprise apps to Next.js (SSR & Server Actions), slashing page load times by over 90%.
—Achieved a 40% increase in overall delivery speed and a 95% reduction in translation time by developing and integrating custom agentic AI workflows, demonstrating real AI ROI.
—Built a highly scalable Next.js/AWS S3 data pipeline utilizing multipart chunked uploads to handle massive 50GB file transfers without timeouts.
—Developed a custom async state management library with timed caching, reducing API polling and optimizing the existing Redux architecture.
—Deployed secure Backend-for-Frontend (BFF) proxy APIs to interface with Java microservices, Apigee, and Salesforce.
—Architected a custom i18n routing system from scratch using advanced Next.js rewrites to bypass strict enterprise deployment constraints.
—Acted as the primary Next.js SME, mentoring 10+ engineers (including senior engineers) on modern React paradigms and resolving architectural blockers.

Jul 2023 — Dec 2023

React (Vite)PostgreSQLExpress.js

—Engineered an end-to-end web app using React, Express, and PostgreSQL, driving the full lifecycle from schema design to deployment.
—Implemented client-side performance optimizations including component lazy loading, dynamic routing, and optimistic state updates.

NULL

Please don't ask me to reverse this linked list 😫.

// TODO: remove debugging log before pushing to master

You can try clicking on that log and see what happens!

Projects

Selected works exploring real-time workflows, interactive interfaces, and scalable infrastructure. Each project built from scratch to solve real bottlenecks and measure real metrics.

01.

Hangout Lobby

Real-Time Chat & Community Platform

ZeroMessage Drop Rate

WebRTCSub-100ms Voice & Video Latency

InfiniteCursor-based Infinite Scroll

WebSocketsWebRTCSocket.ioTanStack QueryNext.js

Case Study

Expand

Problem Statement

Building a low-latency, highly reliable communication platform that handles complex state across text, audio, and video without crashing during connection drops.

01.Bypassed Next.js serverless limitations by custom-configuring a NextApiRequest hack to handle persistent bi-directional WebSockets.
02.Managed complex global state for audio/video permissions and active UI interactions using Zustand and custom React hooks (useChatSocket, useChatScroll).
03.Ensured secure and seamless media sharing utilizing UploadThing for robust file handling.

Project Links & Artifacts

Explore live production builds or view codebases on GitHub.

View Source Code

02.

AI Companion

Conversational AI SaaS Platform

13BModel Parameters

100%Session Memory Retention

ServerlessStripe Webhook Processing

Llama-2 13bLangChainPinecone Vector DBUpstash RedisReplicate

Case Study

Expand

Problem Statement

Providing a scalable, monetized platform that enables non-technical users to easily engineer and interact with custom AI models while maintaining persistent conversational context.

01.Successfully managed LLM context windows by chunking and retrieving conversational history from the vector database before generating responses.
02.Ensured seamless, low-latency UI updates during model inference using custom hooks and streaming responses.
03.Designed a flexible Prisma schema to associate custom companions, chat histories, and Stripe subscription statuses accurately with Clerk user IDs.

Project Links & Artifacts

Explore live production builds or view codebases on GitHub.

View Source Code

03.

Multi-Tenant E-Comm CMS

Headless e-commerce CMS & Decoupled Storefront

IsolatedMulti-Store Data Contexts

DynamicRuntime REST Endpoints

HeadlessDecoupled Client Architecture

Multi-Tenant ArchitectureStripe PaymentsHeadless CMSMySQL (PlanetScale)TanStack Table

Case Study

Expand

Problem Statement

Building a scalable, multi-tenant architecture where a single admin application can securely control data, media, and API endpoints for infinitely many independent storefronts.

01.Architected the initial schema in MySQL via PlanetScale, later executing a full database migration to MongoDB, adapting Prisma relational models to document structures to optimize infrastructure costs.
02.Designed complex relational data models in Prisma to ensure strict data isolation and cascade-deletion integrity between different stores.
03.Managed sophisticated client-side state across two separate repositories using Zustand for shopping carts and admin UI toggles.
04.Implemented an advanced data table system using TanStack Table to handle large inventory datasets with complex filtering and pagination.

Project Links & Artifacts

Explore live production builds or view codebases on GitHub.

Admin Dashboard Repo Storefront Repo

04.

Learnify

Comprehensive Learning Management System (LMS)

HLSAdaptive Bitrate Streaming

RBACRole-Based Access Control

Dual-StoreSanity + Prisma Architecture

Mux (HLS Video)RazorpaySanity CMSNext-Auth v5Uploadthing

Case Study

Expand

Problem Statement

Engineering a scalable educational platform that handles heavy media streaming, complex access control logic, and a multi-tiered affiliate commission system.

01.Managed complex database relations to accurately calculate and persist granular student progress across hundreds of individual video lessons.
02.Handled secure video file uploads and optimized playback delivery through UploadThing while maintaining seamless YouTube embed fallbacks.
03.Upgraded and configured Next-Auth v5 (Auth.js) to manage complex session states across various user roles (Admin, Creator, Student, Affiliate).