Loading…

Hello, I’m Ammar Shaikh. A –

Ammar Shaikh — Full-Stack Developer & UI/UX Designer in Surat, India | Affordable Website & Application Maker — Developer who Ships

Full-Stack Rhythm & Focus

0:00 2:00

I’m Ammar Shaikh, a full-stack software developer in Surat, Gujarat, India. I build React and Next.js interfaces, MERN web applications, and backend APIs for clients and teams.

Three years shipping production software across MERN, Python, and Laravel – delivering 15+ live client and in-house applications at CoderKube Technology (including JD Web and Ship), plus independent commercial builds like CrystalStonesMart.com.

I believe great software combines resilient backend architecture with crisp frontend execution – and I own delivery from database design to the final deployed pixel.

Technologies I Work With

React Node.js Express Python AI / ML LLM SQL MongoDB FastAPI JavaScript TypeScript Git Docker REST APIs

React Node.js Python MongoDB FastAPI LLM React Node.js Python MongoDB FastAPI LLM

Express SQL AI / ML TypeScript Docker Git Express SQL AI / ML TypeScript Docker Git

CrystalStonesMart

B2B wholesale e-commerce platform
with admin analytics & 100+ listings.

Nutrition Mate

Health platform with BMI tracking,
calorie planner & Firestore backend.

CoderKube Systems

15+ production apps (JD Web & Ship)
serving 10,000+ users via REST APIs.

Upwork Retainer

Weekly full-stack delivery slot,
JWT hardening & 30% faster queries.

TESTIMONIALS

In their words

Best experience. Ammar made our productivity and advertising easy with Crystal Stones Mart. Highly efficient, reliable, and consistently delivers quality and perfection on time.

Adjaz Shaikh

Owner at Crystal Stones Mart

Using Ammar's Nutrition Mate I manage health check-ups and custom fitness planning for my personal clients efficiently. Truly helpful for nutrition tracking and client progress.

Faruk Shaikh

Fitness Consultant & Gym Trainer

Ammar has solid knowledge of working with live websites, database schemas, and REST APIs. Great understanding of full-stack development, ownership, and engineering discipline.

Piyush Nakrani

Full-Stack Mentor (React & Node.js)

I design & engineer full-stack web applications. Now I build products teams and clients depend on.

Based in Surat, Gujarat, India with 3+ years in production, I bring end-to-end full-stack development and UI/UX design ownership across the entire software lifecycle. At CoderKube Technology I’ve shipped 15+ live client and in-house projects – from JD Web to Ship – both as solo lead and in Agile sprints across React, Node, Express, MongoDB, Python, and Laravel.

Alongside, I run an active freelance practice on Upwork, Fiverr, and direct retainers, delivering commercial products like CrystalStonesMart.com. Clients trust me for three things: a modern interface designed in Figma and coded in React that feels snappy and polished, a secure REST API with JWT and RBAC that scales smoothly, and disciplined async communication that never needs chasing.

Professional Experience

Full-stack developer shipping production web applications, backend APIs, and scalable databases. Delivered 15+ client and in-house projects across MERN, Python, and Laravel.

Delivered 15+ client-based and in-house production projects (including JD Web and Ship), handling solo builds and team collaborations across MERN, Python, and Laravel.
Engineered 15+ RESTful APIs with JWT authentication and role-based access control (RBAC), consumed by web and mobile frontends serving 10,000+ users.
Optimized MongoDB queries and redesigned collection schemas, improving API response times by approximately 30% across multiple client projects.
Led end-to-end development on solo projects while collaborating in Agile sprints with code reviews and CI/CD pipelines.

Maintaining an active weekly retainer on Upwork, delivering continuous features, bug fixes, and third-party API integrations for overseas clients.
Architected and delivered CrystalStonesMart.com as an independent MERN stack e-commerce build for a paying client: catalog, cart, Google OAuth, orders, and analytics portal.
Built Nutrition Mate (health & wellness platform) utilizing React and Firebase Firestore with BMI calculation and calorie planning.
Available for weekly engineering retainers and project contracts with US, UK, and European timezone overlap.

Completed professional full-stack training covering advanced JavaScript (ES6+), React.js component state patterns, Node.js runtime, and Express middleware.
Implemented complex MongoDB aggregation pipelines, Mongoose schemas, and relational database integrations.
Built production-grade web applications featuring JWT session management, secure password hashing, and responsive UI layouts.

B.E. in Computer Engineering from Gujarat Technological University (GTU). First Class with Distinction – CGPA 7.5+. Graduating April 2026.
Certified AI Machine Learning Engineer – NSQF Level 5 – NASSCOM (Skill India), Aug 2024.
Solid theoretical foundation in data structures, algorithms, operating systems, database management systems, and network security.

DailyStack

Sprint

Sprint Checklist

Ship production release

Optimize Mongo aggregation

Upwork client sprint review

Test new React & AI APIs

Review pull requests

Code rhythm that never gets old... @AmmarShaikh

0:00-2:13

First, solve the problem. Then, write code.

– john johnson

Engineering Scalable REST APIs with Node & Mongo

Read Overview

Shipped live CrystalStonesMart

+100

CrystalStonesMart

B2B Wholesale E-Commerce Platform

CrystalStonesMart is a full-stack B2B wholesale e-commerce platform engineered independently for a paying client. It transforms traditional manual order operations into an automated, high-converting digital storefront.

Built with React.js, Node.js, Express.js, and MongoDB, the system features a dynamic product catalog of 100+ crystal items, shopping cart, custom order flows, Google Authentication, JWT role-based access control (RBAC), and an administrative analytics portal. Delivered end-to-end with mobile-first responsiveness and SEO optimization.

Role

Lead Full-Stack Developer

Client

Crystal Stones Mart

Status

Live in Production

Platform

MERN (React · Node · Express · Mongo)

The showcase app

A wholesale portal that automates every stage of the order lifecycle

Wholesale commerce requires flexible minimum order quantities, real-time inventory adjustments, and friction-free inquiry flows. CrystalStonesMart gives the owner complete oversight while providing buyers an intuitive catalog experience.

Three architectural pillars

BuyerFast product discovery

Explores 100+ crystal items with real-time category filtering, adds items to cart, and signs in instantly using Google OAuth or email.

AdminComplete portal control

Full dashboard for adding, updating, and removing inventory, inspecting buyer orders, and monitoring visitor engagement metrics.

BackendSecure & scalable

Node.js and Express RESTful API with JWT authentication, bcrypt hashing, role-based protection, and indexed MongoDB collections.

9:30

Hatcha

Birthday Potluck Hackathon Anniversary Birthday

The Problem

Every new event type meant new flows and still fell short

Design a wedding planner and you get one set of flows. A corporate retreat, another. A kid’s birthday, a potluck, each different. Pre-designing for every scenario meant the product felt like a generic form with extra steps, and still couldn’t cover the unique edge cases.

What needed to shift

Pre-designed

The GenUI answer

Design a bespoke set of flows for every event type

→

Stop trying to design every screen

A generic form with extra steps that still misses the edges

→

Design the system the AI composes from

“The screen wasn't the unit of design anymore. The goal was. Once I made that shift, everything else followed.”

Mindset shift

I had to stop thinking in user flows

Early on I was still designing screens: host interview flow, dashboard flow, guest response flow. That’s what designers do. But the more we worked with GenUI, the more obvious it became: the screen wasn’t the unit of design. The goal was.

The shift in thinking

Before

GenUI approach

Map every event type, build flows for each — still miss edge cases

→

Define what the user is trying to do; let GenUI compose the right surface

Screen is the unit of design

→

Goal is the unit of design

Fixed output — same UI for everyone

→

Composed at runtime, responds to this user, this event, this moment

What I did

Recognised partway through that I was mapping the wrong thing. Stopped building flows and started mapping goals: what the host needs at each stage, what a guest needs at each moment, then redesigned my deliverables around that intent.

The criteria in action: same data, three different goals

Guests responding

Host dashboard — next day

Host dashboard — dietary breakdown chart

Shopping list — preparing for the event

Core framework

We defined the shell: what stays fixed, what GenUI composes

Users shouldn’t have to relearn the interface for every event. The shell remains fixed, creating predictable navigation and actions, while GenUI customizes everything inside to match the event’s context and intent.

The fixed / flexible boundary

Fixed — every time

GenUI composed — different every time

Navigation and bottom bar

Interview questions, shaped by event type and intent

Progress indicator

Input components: sliders, chips, selects, free text

Primary CTA · close & escape hatches

Dashboard modules · planning cards · guest response surfaces

Page title and shell structure

Background image controls

My contribution

Working with my co-designer, I mapped the fixed/flexible boundary across every surface. My focus was the interview: how the agent’s questions and the components it reaches for shift together with event context.

One shell, two composed interiors — the interview and the event summary

Label

2-3 lines desciption

Low-Key Lounge

High-Energy Rager

Activity-Centric

Social-Centric

Let guests decide

Label

Are we thinking intimate 10-person dinner or a 150-person rager?

50 %

Value 1Value 2

Label

2-3 lines desciption

Place holder text

Helper

Let guests decide

NavigationFixed / Consistent

Interview questionsGenUI composed /
Flexible

Primary CTAFixed / Consistent

Event Details

Name

Maya’s 30th Birthday

Date

Sat, Oct 24

Time

7:00 PM – late

Location

The Loft on 5th

Guests

32 invited

Theme

Retro Disco

Budget

$1,200 all-in

Dress code

’70s glam

Food & Drinks

Style

Family-style dinner

Tacos + mezcal bar

Cake

Chocolate fudge

Drinks

Signature punch

Guest Questions

Drink Questions

Any allergies or a go-to cocktail?

Arrival Time

Can you make it by 7 for the surprise?

Dietary Restrictions

Vegan, gluten-free, or anything to avoid?

Vibe & Music

How the night should feel

Playlist

Disco classics, guest requests open

Lighting

Warm amber + a disco-ball moment

Surprise toast

Champagne at 9:30 PM

NavigationFixed / Consistent

Event summary cardsGenUI composed /
Flexible

Primary CTAFixed / Consistent

My core work

Building the system GenUI composes from

GenUI is only as good as the components it can reach for and the rules that govern them. Before AI could compose interfaces, it needed a design system that balanced flexibility with consistency, and a large part of the work wasn’t designing components but teaching GenUI when and how to use them.

What I did

Built the entire design system from scratch: atoms, tokens, form and GenUI-specific components. I authored the component spec system (use-when / avoid-when / examples) fed into each GenUI loop as system prompts: the foundation for how GenUI selected and composed UI.

1 · Build the system

Atoms → tokens → form components → rich components → data-viz components

2 · Define variability

Spec each component for the range of contexts GenUI will invent

3 · Author criteria

Use-when / avoid-when examples injected as system prompts

4 · Test & refine

See what GenUI reaches for, fill the gaps, rewrite until right

Theme system

Designing the rules for colour and image, not the output

As a host answers the interview, GenUI generates a themed invite, with no palette and no background image drawn in advance. My job was to design the rules that make every generated result feel intentional.

What I did

Designed the colour system and wrote the grading logic and prompt that governs how GenUI selects event-appropriate palettes. Separately defined the image-generation guardrails: mood, composition, positioning, exclusions, so every background feels intentional and brand-safe regardless of event type.

The colour palette system

I designed a four-role colour system. The prompt I authored tells the AI how to select a palette using that grading system as the constraint. A children’s birthday and a corporate dinner will never share a palette, but both will always be internally coherent.

Validate

Accessibility contrast

Dark mode optimized

Cohesive visual identity

Image generation guardrails

I also defined how background images are generated, from mood and composition to visual character, based on insights gathered from the host.

Should include

Mood & visual character suited to event type

Image positioning and compositional rules

Palette relationship with the generated theme

Always excluded

No humans or faces

No text or typography

No logos or brand marks

“Less flow-charting. More system thinking. The designer's job didn't disappear – it moved upstream.”

Outcome

Shipped at Google I/O 2026

Demonstrated live to a developer audience as the showcase for Flutter’s GenUI SDK. One of the first design systems built on the alpha, the component criteria we wrote contributed to how the Flutter team thought about GenUI’s design guidance.

What shipped

Shipped

Demonstrated live at Google I/O 2026 and Google Next 2026 to a global developer audience

0 screens

The dashboard was never fully designed — it was composed at runtime from the system we built

Alpha SDK

One of the first design systems built on Flutter GenUI

New model

The fixed/flexible framework became reusable design thinking for GenUI products

Hatcha event planner running on an iPad Pro

What designing for GenUI
actually changed

The work didn’t disappear. It moved. From drawing every screen to curating a system the AI composes from. From mapping flows to mapping goals. From predicting every scenario to defining the rules of every scenario.

Read next

Fireflut

Designing a Gemini demo for the MWC floor 2025.

Nutrition Mate

Health & Wellness Web Platform

Nutrition Mate is an end-to-end health and wellness tracking platform designed and developed by Ammar Shaikh. It provides fitness enthusiasts and personal trainers with intuitive biometric tracking, calorie goals, and meal planning.

Role

Sole Full-Stack Creator

Client

Health & Wellness Community

Shipped

Live on Vercel

Platform

React.js · Firebase Firestore · Vercel

The Problem

Complex wellness tracking, made radically simple

Most health calculators are cluttered with ads and convoluted forms. Nutrition Mate simplifies daily tracking with instant BMI calculations, calorie-targeted recipe recommendations, supplement store catalog, and direct user health records.

What needed to shift

Before

After

AI buried in navigation

→

AI at the surface, always in view

Text-heavy usage summaries

→

Glanceable gauge – read in two seconds

Light, generic carrier UI

→

Dark, high-contrast, AI-first

No sense of momentum or progress

→

Live activity feed that rewards return visits

Before – the original Fireflut app: cluttered bill and usage summaries, the AI recommendation buried low, a generic light carrier UI, with callouts flagging unclear data viz and low AI visibility

After – the redesigned Fireflut app: a glanceable data gauge, location-based Gemini insights, an elevated AI chat card, and a prominent AI button in the navigation bar

The redesign started with a simple question: would this stop someone walking past the booth? To answer that, we restructured the content hierarchy by bringing the most important data forward and creating a clearer path from curiosity to understanding.

Direction

Dark, high-contrast, AI-first

Carrier apps are light and predictable. Breaking from that made the AI feel like the point – not an add-on. Three principles guided every screen.

Three principles

Dark canvas

Instant contrast on a bright show floor. Depth that makes AI feel alive.

Recognizable at a glance

A consistent blue accent distinguishes Gemini-powered interactions, making AI moments easy to identify throughout the app.

Familiar bones

Plan cards and usage stayed recognisable. New skin, trusted structure.

Moodboard

Color scheme

Background020202

Background glow3456C8 → 000000

Primary3456C8

TextFFFFFF

Secondary4B9BFF

Accent-13700A4

Built for every brand.

The experience was designed with flexible theme variants, allowing telecom providers to seamlessly adapt the interface to their own brand identity while maintaining a consistent AI experience.

Home screen

Glanceable data, AI in reach

A bold usage gauge replaces the wall of text. The Gemini entry point sits high and stays visible – no scrolling to find the smartest part of the app.

Recommendations for you

Consider an international data pass

You've traveled abroad recently, a data pass helps avoid overcharge fees

Upgrade to Wolverine plan

Your data usage is high, this plan gives you more data

Add premium scam shield

Enhanced scam blocking can help you protect from unwanted calls

Chat details

Small motions that make AI feel responsive

The chat's polish lives in its motion. Pre-prompts lower the barrier to the first message; responsive animation makes Gemini feel present and quick.

9:30

‹

Which plan does what I need and costs the least?

9:30

‹

Which plan does what I need and costs the least?

Outcome

Shipped to the MWC 2025 show floor

Designed, built, and brought to production in just 3.5 weeks, then demoed live as Google Cloud's telecom AI showcase.

Results

3.5 wks

Full redesign and build, end to end, no scope cuts

MWC 2025

Demoed live as Google Cloud's telecom AI showcase

Future work

Sparked new telecom AI use cases post-event

The redesigned Fireflut home screen on an iPhone resting on dark fabric — the Gemini-powered telecom assistant showing data usage, plan details, and AI recommendations

Convention was the thing
worth breaking

The redesign leaned into contrast, a dark, AI-forward feel over the familiar carrier look, and let small motions carry the sense that this was something new. That's what made it land on the world stage.

Read next

Existence

A time-intelligence app built for Rob Dyrdek, releasing 2026.

CoderKube Systems

15+ Production Applications (JD Web, Ship)

At CoderKube Technology, Ammar engineered and shipped 15+ live client-based and in-house production web applications across MERN, Python, and Laravel. Built robust, secure RESTful APIs serving 10,000+ active users.

Role

MERN Stack Developer

Company

CoderKube Technology

Scale

15+ Live Apps · 10K+ Users

Platform

React · Node · Express · Mongo · Laravel

CoderKube applications running in production

My core work

What I owned across production builds

From solo project ownership on projects like JD Web and Ship to high-velocity Agile sprints, my responsibilities centered on API resilience, database performance, and end-to-end delivery.

Technical pillars

RESTful API Engineering

Engineered 15+ RESTful APIs with JWT authentication, bcrypt hashing, and role-based access control (RBAC), consumed by web and mobile frontends serving 10,000+ users.

MongoDB Optimization

Redesigned collection schemas and optimized aggregation pipelines, cutting API response times by approximately 30% across multiple client projects.

Cross-Stack Delivery

Delivered high-converting web portals across React.js, Python services, and Laravel backends with clean MVC structures and reusable components.

Agile & Code Reviews

Led solo builds while collaborating in team sprints with git workflows, peer code reviews, and automated deployment pipelines on Vercel.

The outcome

Delivering at scale with proven reliability

Every project shipped to real users and real paying clients, reinforcing a track record of disciplined full-stack execution.

Measured results

15+ Apps Shipped

Client-based and in-house products delivered on schedule (JD Web, Ship, etc.)

10,000+ Users

Actively served by robust JWT-authenticated REST APIs with role-based access

~30% Latency Cut

Through MongoDB collection schema redesign and index optimization

Time Block form

Energy rating

Adding reflections

Reach out to know more

Existence is still under wraps before launch, so there's only so much I can show here. Get in touch and I'll walk you through the deeper work.

Read next

Jumpable

Benchmark your vertical from a jump lab on your ankle.

Upwork Retainer

High-Performance Weekly Engineering Slots

A dedicated freelance practice providing ongoing engineering retainers and project delivery for overseas clients across Upwork, Fiverr, and direct private contracts.

Specializes in rapid full-stack delivery, API hardening, JWT & RBAC authentication, and database refactoring. Comfortable with UK, US, and EU business hours, providing reliable async communication and weekly shipped code.

Role

Senior Full-Stack Freelancer

Platforms

Upwork · Fiverr · Direct Retainers

Availability

Weekly Slots (Global Timezones)

Stack

MERN · REST APIs · Performance

The problem

Shipping features and fixes without stalling the team

Overseas founders often struggle with slow ramp-ups, poor async communication, and brittle code. With a dedicated weekly retainer slot, clients get a full-stack engineer who understands their codebase, owns tickets from start to finish, and delivers production-ready PRs every week.

The impact

What the retainer delivers each week

Predictable velocity, rapid bug fixes, and rock-solid API security. Clients don't have to chase updates or worry about breaking changes.

Measured advantages

24h

Fast response and turnaround time on weekly retainer tickets and urgent production patches.

Large commercial products delivered end-to-end, including CrystalStonesMart.com and Nutrition Mate.

Ownership

Direct scope management, proactive communication, and zero overhead for founders and tech leads.

Jumpable on Apple Watch Ultra The Jumpable app on an Apple Watch Ultra, showing jump height with RSI, flight and ground contact metrics

Jumpable worn on the ankle An athlete mid-jump with the Apple Watch running Jumpable strapped to the ankle

Questions that set the priorities

Who is the experience for?

When a product tries to serve everyone, every workflow compromises. We defined Jumpable's target user, giving every design decision a clear North Star.

How do we help users understand it?

To make an intuitive first impression, we streamlined onboarding and the home screen so athletes immediately understand Jumpable's value.

What is the primary focus?

If the core action isn't obvious, every feature competes for attention. We anchored on one: take a jump, read your number.

Illustration of a Jumpable sensor worn at the ankle above running shoes

Phase 1 · Onboarding

Building onboarding from scratch

Sign-up dropped athletes straight into the app with no context. I created a lean, watch-first onboarding experience with concise microcopy and clear guidance, making the first interaction intuitive and confidence-building.

Phase 2 · Home + brand

Redesigning the post-jump home screen

Post-jump, every metric previously competed for attention with equal visual weight, burying key insights. I restructured the Apple Watch interface around a single focal point, placing the primary jump metric front and centre using a dedicated hero accent colour, while tucking secondary stats into a natural scroll. This clear visual hierarchy allows athletes to instantly scan their core results on a high-motion ankle display.

The numbers were always there.
Now they land

Jumpable could always measure the jump. My work made it something athletes actually understand, trust, and rely on, from their very first launch to a split-second glance after a rep.

Three Apple Watches showing the Jumpable app: jump history, the Meet Jumpable intro, and live jump metrics

Read next

Hatcha

How we prototyped generative UI for Google's biggest developer stages.

Ammar Shaikh — Full-Stack Developer & UI/UX Designer in Surat, India | Affordable Website & Application Maker — FULL-STACK / MERN / UI-UX Developer who Ships

Technologies I Work With

In their words

I design & engineer full-stack web applications. Now I build products teams and clients depend on.

Playground — code and UI experiments by Ammar Shaikh

CrystalStonesMart

A wholesale portal that automates every stage of the order lifecycle

Every new event type meant new flows and still fell short

I had to stop thinking in user flows

We defined the shell: what stays fixed, what GenUI composes

Building the system GenUI composes from

Designing the rules for colour and image, not the output

Shipped at Google I/O 2026

What designing for GenUIactually changed

Fireflut

Nutrition Mate

Complex wellness tracking, made radically simple

Dark, high-contrast, AI-first

Built for every brand.

Glanceable data, AI in reach

Small motions that make AI feel responsive

Shipped to the MWC 2025 show floor

Convention was the thingworth breaking

Existence

CoderKube Systems

What I owned across production builds

Delivering at scale with proven reliability

Reach out to know more

Jumpable

Upwork Retainer

Shipping features and fixes without stalling the team

What the retainer delivers each week

Building onboarding from scratch

Redesigning the post-jump home screen

The numbers were always there. Now they land

Hatcha

Ammar Shaikh — Full-Stack Developer & UI/UX Designer in Surat, India | Affordable Website & Application Maker — Developer who Ships

What designing for GenUI
actually changed

Convention was the thing
worth breaking

The numbers were always there.
Now they land