Ahmad HassanFull-Stack AI Engineer
Hire me

Full-Stack AI Engineer in Lahore, Pakistan · AI · Web · Android

I build AI products
that survive real users.

I’m Ahmad Hassan, a full-stack engineer and AI engineer in Lahore. Most LLM work stops at “it calls the model and runs on my machine.” Mine is engineered for the parts that break in production — token budgets, tenant isolation, revocation-aware sessions, and deploys that actually happen.

See the work Get in touch
Products shipped
11
Automated tests written
1,685
Families on EasyQuran
10,000+
Google Play rating
4.9★
Ahmad Hassan as a stylised 3D character in a denim jacket, standing in front of his coding setup
The workstation plays a short loop: code, think, debug, fix, test and ship. Jump to a step:
Google GeminiOpenAILangChainRAG pipelinesTool callingVector searchStreamed outputPandasReact 19Next.js 15 / 16Server ComponentsTypeScript (strict)Tiptapshadcn/uid3-forceNode.jsNestJSExpress 5FlaskServer-Sent EventsSocket.IOBullMQZodPostgreSQLPrismaSupabaseRow Level SecurityMongoDBAtlas Vector SearchFirestoreRedisKotlinJavaJetpack ComposeMaterial 3WorkManagerReact NativeGitHub ActionsDockerVercelCloudflare R2StripeVitest

01Selected projects by Ahmad Hassan

Products built to survive real users.

Eleven products across AI, web and mobile. Open any card for the full case study — every number comes straight from the repo.

EasyQuran case study
EasyQuran — Native study app01

Native study app

EasyQuran

The Quran app for every Muslim family.

families
10,000+
Google Play
4.9★
translations
11+
platforms
3

MobileProduct

Aria case study

need an appointment tomorrow after 5

tool availability · e-Booking service

These slots are open tomorrow — shall I hold one?

17:3018:0018:45

otp verification sent · audit logged

02

AI booking assistant · Prime Coworking

Aria

An AI booking assistant that owns none of the booking logic.

tool-calling functions
15
message bounded history
40

AIWeb

Retrivo Vault case study
Retrivo Vault — AI research assistant · RAG SaaS03

AI research assistant · RAG SaaS

Retrivo Vault

The research assistant that only knows what you’ve read.

tests
142
chunking cascade
3-tier

AIWeb

LeadForge AI case study
LeadForge AI — AI sales-intelligence platform04

AI sales-intelligence platform

LeadForge AI

Find the right businesses. Understand them. Start the right conversation.

automated checks
198
scoring dimensions
6

AIWeb

RetailFlow case study

Bill #1042Shift · Counter 1

چاولBasmati rice · 2.5 kgRs 1,125

چینیSugar · 1 kgRs 165

دودھMilk pack · ×3Rs 810

TotalRs 2,100

Cash 1,500Khata 600
05

Multi-tenant retail SaaS

RetailFlow

Retail software for shops that trade on khata.

tests
231
tables
35

WebProduct

NoteMind case study
NoteMind — AI knowledge platform06

AI knowledge platform

NoteMind

Notes that think with you.

tests
173
vulnerabilities
0

AIWeb

FitMind AI case study

Plan change proposal

Your last three sessions were rated brutal.

Reduce weekly volume and keep your goal date?

ApproveKeep plan
07

Adaptive AI fitness coaching

FitMind AI

A coach that re-plans around you — and asks before it changes anything.

tests
941
decision engines
8

AIMobile

The Copper Larder case study
The Copper Larder — Streaming AI restaurant host08

Streaming AI restaurant host

The Copper Larder

Hannah — an AI front-of-house concierge for hospitality sites.

stage request pipeline
8
zero-token intercepts
~19

AIWeb

MindScribe case study
MindScribe — Privacy-first Android journal09

Privacy-first Android journal

MindScribe

Journaling that treats your words as sensitive data.

classes
25
lines of Java
3,857

MobileProduct

SpendSmart case study
SpendSmart — Real-time Android expense tracker10

Real-time Android expense tracker

SpendSmart

Every edit lands on every device before the keyboard closes.

backend servers
0
dependency lines
8

MobileProduct

FitTrack case study
FitTrack — Gym management dashboard11

Gym management dashboard

FitTrack

Gym management, without the spreadsheet.

routes
8
UI dependencies
0

WebProduct

02What I build

Web, AI and Android products, built end to end.

Based in Lahore, working with teams and clients locally and remotely.

Web applications and interfaces

Responsive React and Next.js front ends on a hand-built design system, backed by NestJS, Node or Python APIs and PostgreSQL, MongoDB or Supabase.

Shipped inRetailFlowFitTrackNoteMind

AI products

Retrieval over private documents, tool-calling assistants on live services, streamed answers with citations, and the cost, quota and safety controls a public LLM endpoint needs.

Shipped inAriaRetrivo VaultLeadForge AI

Android apps

Native Kotlin and Java apps with offline-first sync, encrypted storage, Firebase and Play Store releases.

Shipped inEasyQuranMindScribeSpendSmart

Backends and multi-tenant SaaS

Secure APIs, tenant isolation in the database, Stripe billing, queues and background jobs, shipped with CI.

Shipped inRetailFlowRetrivo VaultLeadForge AI

03Engineering judgement

Decisions I’d defend in a review.

The calls that separate a demo from a product. Pick one to see where it shipped.

01

Money is never a float.

RetailFlow stores every amount as an integer and reconciles the cash drawer to the paisa before anyone goes home.

Shipped inRetailFlowRead the case study →

04How I build

Seven rules, and the tools I reach for.

  1. 01

    Validate at the boundary.

    Every write goes through a schema; bad input never reaches business logic.

  2. 02

    Fail fast, at boot.

    Misconfiguration should kill the process at startup with a readable error — not surface as a 500 next Tuesday.

  3. 03

    Assume the model misbehaves.

    Rate limits, quotas, timeouts, caches and fallbacks are part of the feature, not a follow-up ticket.

  4. 04

    Isolate tenants in the database.

    Authorization in application code is a suggestion; enforced by Row Level Security it is a guarantee.

  5. 05

    Tests are how I move fast.

    They aren’t bureaucracy — they’re what lets me refactor auth without fear.

  6. 06

    Own the whole path.

    Schema to store listing. If there’s a handoff, something falls through it.

  7. 07

    Ship it.

    Code that isn’t deployed doesn’t count.

~/stackshipped in production

AI & DataGoogle GeminiOpenAILangChainRAG pipelinesTool callingVector searchStreamed outputPandas
FrontendReact 19Next.js 15 / 16Server ComponentsTypeScript (strict)Tiptapshadcn/uid3-force
BackendNode.jsNestJSExpress 5FlaskServer-Sent EventsSocket.IOBullMQZod
DatabasesPostgreSQLPrismaSupabaseRow Level SecurityMongoDBAtlas Vector SearchFirestoreRedis
MobileKotlinJavaJetpack ComposeMaterial 3WorkManagerReact Native
DevOpsGitHub ActionsDockerVercelCloudflare R2StripeVitest

05Experience

Shipped with teams, and alone.

  1. Co-lead developer

    EasyQuran

    One of two lead developers on a Saudi engineering team, alongside dedicated QA, content and marketing. Frontend and backend for a Quran study app live on the App Store and Google Play — 10,000+ families, 4.9★.

    View the work

  2. Product engineer

    Prime Coworking

    Built Aria, the AI booking assistant on top of the company’s e-Booking SaaS — LLM tool-calling over live domain services, OTP-gated mutations and an audit trail.

    View the work

  3. Independent engineer

    Open source

    Nine production-grade products shipped end to end — RAG, agents, multi-tenant SaaS and native Android — every number pulled straight from the repo.

06Questions

Common questions, answered.

Who is Ahmad Hassan?

Ahmad Hassan (GitHub: Ahmadhsn1) is a full-stack AI engineer based in Lahore, Pakistan. He builds AI products, web applications and native Android apps end to end, and is co-lead developer of EasyQuran, a Quran study app used by more than 10,000 families with a 4.9-star rating on Google Play.

What kind of software does Ahmad Hassan build?

AI products such as retrieval-augmented generation (RAG) assistants and tool-calling agents, full-stack web applications with React, Next.js and NestJS, multi-tenant SaaS on PostgreSQL and Supabase, and native Android apps in Kotlin and Java.

Is Ahmad Hassan available for work in Lahore?

Yes. He is based in Lahore, Pakistan, and is open to engineering roles and project work, both locally and remotely. The fastest way to reach him is by email or WhatsApp.

What are Ahmad Hassan’s best-known projects?

EasyQuran, a Quran study app used by more than 10,000 families; Retrivo Vault, a RAG research assistant that cites its sources; LeadForge AI, an AI sales-intelligence platform with 198 automated checks; RetailFlow, a multi-tenant retail SaaS with 231 tests; NoteMind, an AI knowledge platform with 173 tests; and Aria, an AI booking assistant built at Prime Coworking.

Which technologies does he work with?

TypeScript, React, Next.js, Node.js, NestJS, Python and Flask on the web and backend side; PostgreSQL, MongoDB, Supabase, Redis and Firestore for data; Kotlin, Java and Jetpack Compose for Android; and Google Gemini, vector search and LangChain for AI.

Where can I see his work and code?

Eleven projects are documented with screenshots, metrics and engineering decisions on this site, and the source is on GitHub at github.com/Ahmadhsn1. Every number comes straight from the repositories.

How can I contact Ahmad Hassan?

Email ahmad.hsn0099@gmail.com, message him on WhatsApp at +923256522522, or connect on LinkedIn.

07Contact

Have a product
worth shipping?

Open to AI and full-stack engineering roles — remote, from Lahore. Tell me what you’re building.

Email me

Available now · Lahore, Pakistan · PKT