Skip to content

Shahnihal RahmanApplied AI & Digital Product Builder

I Build Digital Products With AI, Code & Product Thinking.

I'm Shahnihal Rahman — an applied digital product builder working across web applications, AI-assisted development, UX, automation, analytics, and digital growth.

India · Open to UAE / Gulf Opportunities

Drag to rotate the system. Tap any part to see what it connects to.

  • Build
  • Design
  • Automate
  • Deploy
  • Optimize

Web Applications · Applied AI · Digital Products · UX · Automation · AI-Assisted Development

The Lab · Currently Building
01What I actually do

From Idea → Working Product

I turn ideas into working digital products using AI-assisted development, modern web technologies, product thinking, and automation.

Problem

01 / 08

Understand the user's and business's actual need.

  1. 01

    Problem

    Understand the user's and business's actual need.

    Before features, the question: what breaks today, and for whom?

  2. 02

    Product Thinking

    Define features, workflows, priorities and user journeys.

    Scope becomes a sequence. What ships first, and what can wait.

  3. 03

    UX / Interface

    Structure interfaces, flows, information architecture and interactions.

    Hierarchy, states, empty cases, error cases. The unglamorous half.

  4. 04

    AI-Assisted Build

    Use Claude, ChatGPT and other AI development tools to accelerate implementation and iteration.

    AI compresses the typing. The architecture stays a human decision.

  5. 05

    Integrations

    Connect APIs, databases, analytics, authentication and third-party services.

    Payments, auth, notifications, data. Where products usually leak.

  6. 06

    Deploy

    Use modern deployment infrastructure such as Vercel and cloud services.

    Shipped is a state, not an event. Environments, config, rollbacks.

  7. 07

    Test

    Debug, validate responsiveness, identify issues and iterate.

    Real devices, real edge cases, real failure paths.

  8. 08

    Improve

    Use feedback, analytics and real user behaviour to improve the product.

    What people actually do beats what we assumed they would.

Build fast. Think deeply. Iterate relentlessly.

02Featured products

Things I’ve Actually Built

Web products first, then the applied AI products. Daktarji and Truepost are real codebases and the screens shown are captured from the real builds, with the technology read out of the repository rather than recalled from memory. The rest say plainly where they are: what is still in development, which stacks have not been read from a repository yet, and which stages of an AI workflow still pass through a human.

01Live · Product in Active Development

Daktarji

Healthcare Discovery · Digital Product · Web Application

Find the right doctor. See a real appointment time. Skip the phone call.

daktarji.com
Discover

The entry point — a patient states what they need and books a real time, before any doctor list appears.

Captured from the live site at daktarji.com, under its own Daktarji branding. Used unretouched: the only edit was trimming the browser chrome and the scrollbar column. Select the screen to inspect it at full size.

  1. The problem

    Finding care usually means calling around, guessing who is available, and trusting an unverified listing. Patients cannot see a real appointment time before they commit.

  2. What it is

    Daktarji is a healthcare discovery and appointment platform designed around how patients actually discover doctors, clinics and healthcare services.

  3. Patient journey

    1. Discover
    2. Explore
    3. Compare
    4. Select
    5. Book

  4. My role

    • Product Concept
    • UX & Information Architecture
    • Web Application Development
    • AI-Assisted Development

  5. Technology

    Stack read from repo
    • Next.js (App Router)
    • React
    • TypeScript
    • Tailwind CSS
    • shadcn/ui
    • Radix UI
    • TanStack Query
    • Zustand
    • React Hook Form
    • Zod

    Stack read directly from the project repository, not from memory.

  6. AI layer

    • Claude
    • ChatGPT
    • AI-assisted implementation
    • AI-assisted debugging
    • AI-assisted documentation
Live Site
02Custom Build · Active Development

Truepost India

Custom Web Application · Digital Product · Media Platform

A media platform rebuilt from a CMS install into a product I own end to end.

Truepost India · custom build
Custom web application

Reader home — the custom application that replaced the CMS front end.

Captured from the custom application now serving readers, not the legacy CMS it replaced. Select the screen to inspect it at full size.

  1. The problem

    An off-the-shelf CMS publishes fine but caps everything else: content modelling, editorial roles, reader accounts and performance all hit a ceiling you cannot move.

  2. What it is

    I evolved Truepost India from its earlier CMS-based platform into a custom-coded web application, taking responsibility across product structure, user experience, development, deployment and continuous iteration.

  3. Product evolution

    1. Earlier CMS platform
    2. Product rethink
    3. Custom content model
    4. Custom web application
    5. Continuous iteration

  4. My role

    • Product
    • UX & Information Architecture
    • Web Application Development
    • Platform Management

  5. Technology

    Stack read from repo
    • Next.js (App Router)
    • React
    • TypeScript
    • Tailwind CSS
    • Framer Motion
    • Lenis smooth scroll
    • three.js
    • React Three Fiber
    • Supabase
    • PostgreSQL

    Stack read directly from the project repository, not from memory.

  6. AI layer

    • Claude
    • ChatGPT
    • AI-assisted implementation
    • AI-assisted debugging
Live Site
03In Development

E-Commerce Web Application

Digital Commerce · Web Application

A commerce product being built from the buying journey backwards.

E-commerce app · in development
Buying journeyPlanned
VariantSizeStock
UIAPIDBPaymentsOrders

Interface preview · structure rebuilt for this page

01 / 05

  1. The problem

    Hosted commerce platforms make storefronts easy and leave the deciding parts rigid: how products are found, how variants and stock are represented, and how an order behaves once money is involved.

  2. What it is

    A custom commerce application currently in development. Scope and architecture are defined; the build is in progress, so everything below is marked honestly as built or planned.

  3. My role

    • Product Definition
    • UX Flows
    • Application Development
    • AI-Assisted Development

  4. Technology

    Planned stack
    • React
    • Next.js
    • TypeScript
    • Tailwind CSS
    • REST APIs
    • JSON
    • Hosted database
    • Authentication
    • Payments provider
    • Analytics

    Planned stack. Nothing here is presented as shipped — this section will be regenerated from the repository once the build is further along.

  5. AI layer

    • Claude
    • ChatGPT
    • AI-assisted implementation

Links will appear here when there is something real to link to.

04Applied · In Active Development

Applied AI Products

Applied AI · Agentic Systems · Workflow Automation

Practical AI products built around real business problems.

Applied AI · workflow shape
  1. 1Trigger
  2. 2AI Agent
  3. 3Tools
  4. 4Data
  5. 5Action
BuildDebugDocsResearchContentOps

Interface preview · structure rebuilt for this page

01 / 05

  1. The problem

    AI tooling is easy to talk about and easy to misuse. The useful question is which parts of real work it can actually take over, and which parts must stay a human decision.

  2. What it is

    Real systems I've built to solve practical problems — an AI search-visibility platform and an agentic influencer-marketing system — alongside the AI-assisted process behind everything else on this page.

  3. My role

    • Applied AI
    • Automation Design
    • Workflow Engineering

  4. Technology

    Practices, not products
    • Claude
    • ChatGPT
    • Gemini
    • Cursor
    • Kiro
    • Product development
    • Debugging
    • Documentation
    • Research
    • Content operations

    Models and build environments rather than a product stack. The two products below carry their own technology lists.

  5. AI layer

    • Claude
    • ChatGPT
    • Gemini
    • Cursor
    • Kiro
05In Development

AI GEO Platform

Generative Engine Optimization · AI Search · SEO

SEO asks how a search engine ranks you. GEO asks how an AI describes you.

AI GEO · visibility test
  1. Prompt
  2. AI model
  3. Response
  4. Competitor set
  5. Gap detection
  6. Recommendation

Interface preview · structure rebuilt for this page

01 / 06

  1. The problem

    When someone asks ChatGPT or Gemini for a recommendation, the answer is generated rather than ranked. A business has no obvious way to see how it is being characterised in that answer, which competitors are named beside it, or whether it appears at all.

  2. What it is

    A platform designed to help businesses improve how their content and brand are discovered and surfaced by generative AI systems such as ChatGPT and Gemini.

  3. Product workflow

    1. Prompt Testing
    2. AI Response Analysis
    3. Competitor Analysis
    4. Visibility Gaps
    5. Recommendations
    6. Optimization
    7. Re-test

  4. My role

    • Product Concept
    • Product Design
    • AI-Assisted Development
    • Web Application Development
    • Testing

  5. Technology

    Stack not yet read from repo
    • ChatGPT
    • Gemini
    • Other LLMs
    • Web application
    • AI-assisted development

    Declared rather than read from a repository. The codebase was not available in this environment, so this lists only what the product is known to work with — the models it tests against, and the fact that it is a web application. The framework and data layer get filled in from the repository the same way Daktarji and Truepost were, rather than guessed at now.

  6. AI layer

    • ChatGPT
    • Gemini
    • Other LLMs
    • AI-assisted implementation

Helps identify opportunities to improve visibility and representation across AI-generated search experiences. It does not promise rankings, placement or citations inside AI answers — no tool honestly can. Nothing is linked here yet; a link goes in when there is a public one to give.

06In Development

AI Influencer Marketing Agent

Agentic AI · Marketing Automation · Workflow Automation

A campaign brief goes in. A qualified creator conversation comes out.

Influencer agent · campaign run
Agentic workflowHuman-in-the-loop
BriefAgent

Interface preview · structure rebuilt for this page

01 / 06

  1. The problem

    Influencer marketing runs on manual searching, spreadsheet shortlists and copy-pasted outreach. The matching is subjective, the work is repetitive, and most of the follow-up never happens.

  2. What it is

    An agentic AI system designed to automate the influencer-marketing workflow from identifying suitable creators to outreach and communication.

  3. Core workflow

    1. Campaign Brief
    2. Creator Discovery
    3. Creator Analysis
    4. Best-Match Scoring
    5. Personalised Outreach
    6. Conversation
    7. Follow-up
    8. Qualified Influencer

  4. My role

    • Product Concept
    • Agentic Workflow Design
    • AI-Assisted Development
    • Automation
    • Web Application

  5. Technology

    Stack not yet read from repo
    • Agentic workflow
    • Campaign-brief reasoning
    • Fit scoring
    • Web application
    • AI-assisted development
    • Workflow automation

    Declared rather than read from a repository. The codebase was not available in this environment, so the list stays at the level that is actually known — the agent stages, and that it is operated through a web application. Models, data sources and platform integrations are left off until they can be confirmed from the repository.

  6. AI layer

    • Agentic workflow
    • Creator analysis
    • Outreach generation
    • AI-assisted implementation

An internal system rather than a public product, so there is nothing to link yet. The scope above describes what runs today, and the stages that still pass through me are marked human-in-the-loop rather than presented as autonomous.

07Academic Project · 2023

Handwritten Digit Recognition

Machine Learning · KNN · Computer Vision

Where the ML foundation started — a classifier built and evaluated from first principles.

Dissertation · 30 pages
B.Tech · 2023

The original B.Tech dissertation submitted to Techno India University, Kolkata, May 2023. Unmodified.

Select any page to read it in place — page navigation, zoom, fit and fullscreen included.

  1. The problem

    Reading a handwritten digit is trivial for a person and non-trivial for a program: the same digit varies in stroke, slant and thickness every time it is written, so the classifier has to work from similarity rather than rules.

  2. What it is

    A handwritten digit classifier built on the k-nearest neighbours algorithm — my B.Tech dissertation at Techno India University, covering the full pipeline from preprocessing through parameter tuning to evaluation.

  3. Processing pipeline

    1. Load Dataset
    2. Preprocessing
    3. Train-Test Split
    4. Feature Extraction
    5. KNN Model
    6. Prediction
    7. Evaluation
    8. Visualization

  4. My role

    • Problem Definition
    • ML Methodology
    • Python Implementation
    • Model Evaluation
    • Technical Writing

  5. Technology

    Stack not yet read from repo
    • Python
    • scikit-learn
    • NumPy
    • OpenCV
    • Pandas
    • Matplotlib
    • Jupyter Notebook

    Read from the project's own dissertation: the libraries section names scikit-learn, NumPy, Matplotlib, Pandas and OpenCV, and the captured output shows the work running in a Jupyter notebook. Nothing is added beyond what the document shows.

GitHub
03AI-assisted development

I Don’t Just Use AI. I Build With It.

From prompt to product.

Back to idea
My call

Idea

The problem worth solving, framed as an outcome.

AI accelerates my development process, but product decisions, architecture choices, testing, validation, quality and final outcomes remain my responsibility.

Tools in the loop

  • ClaudePrimary build & reasoning assistant
  • ChatGPTExploration, drafting, alternatives
  • GeminiResearch & cross-checking
  • CursorIn-editor implementation
  • KiroAgentic build environment — used to build this site
AI-assisted buildAI-assisted debuggingAI-assisted documentationNo fabricated AI expertise
04Interactive architecture

How I Build

The same shape underneath every product on this page. Open a layer to see what actually lives there.

  4. Transactional booking, slot holds, commissions, publishing states, access roles. This runs server-side because correctness matters more than convenience.

    • Transactions
    • Slot holds
    • Role rules
    • State machines

05Technology constellation

The Stack, Mapped Honestly

No skill meters. A bar claiming near-mastery of React tells you nothing you can check. Instead, every technology here carries where it comes from: a real repository, delivered work, or something I am openly still exploring.

Frontend

Where the product becomes something a person can actually use.

  • React
  • Next.js
  • TypeScript
  • JavaScript
  • HTML5
  • CSS3
  • Tailwind CSS
  • Responsive Design
  • Framer Motion
  • three.js

Web Platforms

Years of shipping on hosted platforms before building custom.

  • WordPress
  • Shopify

APIs / Data

The layer that decides whether a product can be trusted.

  • REST APIs
  • JSON
  • Firebase
  • Cloud Firestore
  • Supabase
  • PostgreSQL
  • Row Level Security
  • Zod
  • TanStack Query

Cloud / Deployment

Shipping is part of building, not a separate job.

  • Vercel
  • Firebase App Hosting
  • Cloud Functions
  • Google Cloud
  • Git
  • GitHub
  • Turborepo

AI

Used as a build layer, with the decisions staying mine.

  • Claude
  • ChatGPT
  • Gemini
  • Cursor
  • Kiro
  • AI agents
  • Workflow automation

Analytics / Growth

A product that nobody finds or completes is not finished.

  • Google Analytics 4
  • Search Console
  • Meta Pixel
  • SEO
  • CRO
  • Meta Ads
  • Google Ads
  • Vercel Speed Insights
Evidence key
  • In a project repoPresent in a codebase inspected while building this portfolio.
  • Used in delivered workUsed on live platforms, stores or campaigns I have run.
  • ExploringActively learning or trialling. Listed without claiming depth.

Progress bars on a skills list are a design decision that hides information. A tag with a source behind it tells you more in less space — and it is checkable.

06Product thinking

Technology Is Only Half the Job.

Working code is the entry fee. Whether a product is worth using is decided in the decisions around it.

How the layers pull on each other

  1. Business

    Outcome, constraints, priorities

  2. Product

    Scope, workflows, sequencing

  3. UX

    Structure, flows, interaction

  4. Technology

    Architecture, data, delivery

  5. AI

    Acceleration and automation

A decision at any layer changes the ones above and below it. Treating them as separate is how products end up technically fine and commercially useless.

  • 01

    User experience

    Whether the thing feels obvious in the hand.

  • 02

    Product logic

    What the system should do, and refuse to do.

  • 03

    Business requirements

    The outcome someone is paying for.

  • 04

    Information architecture

    How content and actions are organised.

  • 05

    Visual hierarchy

    What the eye reaches first, second, never.

  • 06

    Performance

    Speed is a feature, and usually the first one.

  • 07

    Accessibility

    Keyboard, contrast, semantics, reduced motion.

  • 08

    Conversion

    Where intent turns into a completed action.

  • 09

    Communication

    Making the technical legible to everyone involved.

  • 10

    Iteration

    Version two is where products get good.

07Communication & collaboration

I Bridge Technology & Business.

I work comfortably with founders, designers, content teams, marketers, business stakeholders and technical collaborators. I focus on making technical concepts understandable, turning requirements into actionable product decisions, and keeping work moving across teams.

How work actually moves

  1. Requirement
  2. Alignment
  3. Design
  4. Build
  5. Feedback
  6. Delivery

Who I work with, and what we exchange

Most delivery problems I have seen were translation problems, not technical ones.

08Experience

Where I’ve Done the Work

  1. Truepost India Pvt. Ltd.

    Apr 2020 – Jan 2023 · Feb 2025 – Present

    Web Application & Digital Platform Lead

    • Custom web application
    • Product development
    • Platform ownership
    • UX
    • Stakeholder communication
    • Analytics
    • SEO
    • AI-assisted workflows

  2. QUL

    Apr 2023 – Feb 2025

    Web & E-commerce Platform Lead

    India / UAE / Thailand / China

    • Shopify
    • E-commerce
    • UI/UX
    • Customer journeys
    • SEO
    • Analytics
    • Digital growth
    • Multi-market coordination

  3. DML Research

    Jul 2020 – Aug 2020

    Web Developer Intern

    • University website
    • HTML
    • CSS
    • JavaScript
    • PHP
    • CMS
    • Requirements
    • Delivery

Achievements

Technical engineering

  • Full-Stack Products Shipped

    Daktarji and Truepost India — built and deployed end-to-end: frontend, backend, database and integrations, both live in production.

  • Agentic AI Workflows

    Multi-stage agents for AI search-visibility analysis and influencer outreach — outreach and conversation stages kept human-in-the-loop.

  • Backend & Data Architecture

    Firestore security rules on one product, Postgres row-level security and 24 versioned migrations on the other.

  • AI-Assisted Engineering

    Claude, ChatGPT, Gemini, Cursor and Kiro used across AI-assisted development workflows for implementation, debugging and documentation — architecture and final decisions stay mine.

Product & business impact

  • 2.5M+

    Monthly organic reach achieved at Truepost

  • 13K+

    Followers built at Truepost

  • ₹0.033

    Meta campaign CPE achieved

  • IIT Delhi

    Entrepreneurship Development Program

09About

Between Technology & Business

I'm a Computer Science graduate with 5+ years of hands-on experience building digital products, web applications, e-commerce experiences and growth systems. My work sits between technology and business — I enjoy taking ambiguous ideas, structuring them into usable products, and using AI-assisted development to move from concept to working software quickly.

I work comfortably with technical and non-technical stakeholders and enjoy the complete product journey: understanding the problem, designing the experience, building the solution, deploying it, and continuously improving it.

From idea to interface.

Education

  1. Indian Institute of Technology, IIT Delhi

    Entrepreneurship Development Program

    2025 – 2026

  2. Techno India University, Kolkata

    B.Tech — Computer Science Engineering

    2019 – 2023

Shahnihal Rahman

Shahnihal Rahman

Applied AI · Digital Products

B.Tech CSE
Techno India University
Entrepreneurship Development Program
IIT Delhi
5+ Years
Digital / Web Experience
Applied AI
AI-assisted Development & Automation
Contact

Have an idea worth building?

Let’s turn it into something real.

Email MeLinkedInGitHub
Email
rahman@shahnihal.com
Phone
+91 8910404283
Location
India · Open to UAE / Gulf Opportunities