Available for new projects

Marwan
Khan.

Full-Stack Developer & AI/ML Builder

I build practical web applications and explore AI to turn ideas into useful products — scoped cleanly, architected properly, and shipped to production.

View ProjectsContact Me
GitHubLinkedInRésumé
01
product in production
20+
tools & technologies
24 / 7
learning mode
route.tsschema.prisma
// barhtafit — app/api/workouts/route.ts
import { auth } from "@/lib/auth";
import { db } from "@/lib/db";
 
export async function POST(req: Request) {
  const session = await auth();
  if (!session?.user) {
    return Response.json({ error: "Unauthorized" }, { status: 401 });
  }
  const { exercises, duration } = await req.json();
 
  const workout = await db.workout.create({
    data: {
      userId: session.user.id,
      exercises,
      duration,
      loggedAt: new Date(),
    },
  });
 
  return Response.json({ workout });
● TypeScript · Next.js 16Ln 21, Col 32
production
barhtafit.com
build48s
edge · status200 OK
deployed3d ago
ml · training runv0.3
loss
0.124
accuracy
91.2%
epoch 24 / 5048%
01 — About

I build things that ship — and keep shipping.

I'm Marwan — a developer who cares less about tutorial trophies and more about working software. My focus is simple: build real products, learn through the process, and push one level further with every release.

I started with the fundamentals of the web, went deep on React, Next.js and TypeScript, and now design production-grade applications end-to-end. Lately, I'm expanding into AI and machine learning — with one goal: products that don't just work, but get smarter.

Full-Stack Development

End-to-end web apps — from database schema to deployed, polished frontends.

AI & Machine Learning

Studying ML fundamentals with Python, moving from theory to trained models.

Real-World Products

Not tutorials — actual products with users, deployments and iterations.

Learning by Building

Every skill on this site was earned by shipping something with it.

The progressionin motion
Phase 01

Web Foundations

HTML, CSS, JavaScript — understanding how the web actually works.

Phase 02

Full-Stack Applications

React, Next.js, Node and databases — wiring complete systems together.

Phase 03

Product Builder

BarhtaFit live in production; a CRM architected like a real SaaS.

Phase 04you are here

AI / ML Engineer

Currently here — ML fundamentals, Python, first trained models.

Phase 05

AI-Native Products

Next — intelligence built into the products I already ship.

02 — Skills & Tools

A stack built for shipping products.

Every tool here was learned in the field — inside real codebases, real deployments and real deadlines. Dots mark what runs my production projects daily.

/01

Frontend

Interfaces people enjoy using

HTMLCSSJavaScriptTypeScriptReactNext.jsTailwind CSSshadcn/ui
8 tools● active
/02

Backend & Database

Systems that hold up in production

Node.jsMongoDBPostgreSQLPrismaREST APIsAuthentication
6 tools● active
/03

Deployment & Tools

Ship early, ship often

GitGitHubVercelRailwayNetlify
5 tools● active
/04

AI / ML

The current obsession — growing fast

PythonNumPyPandasMachine LearningNLPComputer Vision
6 tools● active
03 — Featured Projects

Products, not just projects.

Each one starts with a real problem, gets architected like a product, and ends up deployed — not sitting in a private repo.

01 — Flagship productbarhtafit.com — live

BarhtaFit

A fitness platform that grew from a side experiment called “Workout Web” into a full product — workout planning, session logging and progress tracking, built with a product mindset from day one.

ProblemMost workout trackers feel like spreadsheets — effort goes in, motivation never comes out.
SolutionPrograms, logging and streaks designed around one loop: train, see progress, come back tomorrow.
ResultLive in production at barhtafit.com — iterated from prototype into a real, evolving product.
Workout builderSession loggingProgress analyticsStreaks & goalsAuth & profiles
Next.jsTypeScriptTailwind CSSPostgreSQLPrisma
Live Demo
BarhtaFit workout library with a personalised workout routine
BarhtaFit training progress analytics dashboard
03 — Reserved slotin R&DAn AI-native product is in development.Machine intelligence meets what I already build — starting with an intelligent engine for BarhtaFit. This portfolio grows with every ship.
04 — The Journey

Building in public, one commit at a time.

No inflated titles, no fake years of experience. Just a clear trajectory — and every step backed by something that actually shipped or got learned.

Web Development Foundations

completed

Started from zero: HTML, CSS and JavaScript — building pages, breaking layouts, learning how the browser thinks.

HTMLCSSJavaScript

Full-Stack Applications

completed

Went end-to-end: React and Next.js on the front, Node and databases on the back, wired together with APIs and auth.

ReactNext.jsNode.jsMongoDB

Production-Oriented Projects

shipped

Moved from 'it works locally' to 'it works for users'. BarhtaFit is live, with workout planning, session logging and progress tracking available to real users.

TypeScriptPostgreSQLPrismaVercel

AI / Machine Learning

in progress

Deep in ML fundamentals right now: Python, NumPy and pandas for data work, then models — regression, classification, evaluation, iteration.

PythonNumPyPandasML theory

AI-Native Product Development

next

The destination: intelligent features inside real products. Recommendation, prediction and vision — applied, not just studied.

NLPComputer VisionApplied AI
marwan@build — zsh
➜ ~/dev git log --oneline --graph
*9f3a2c1(HEAD -> main)ship: barhtafit v2.0 — programs engine
*a8123fflearn(ml): linear regression from scratch
*7c1ad90db: prisma schema — 9 models, full relations
*02b9f88feat(fit): session logging + streaks
*f5c2e01init: next.js + typescript
➜ ~/dev
currently learning
ML math, NumPy & first models
next ship
AI engine for BarhtaFit
status: researching…
05— What's Next

From building software
to building intelligence.

Full-stack development taught me how to turn ideas into products. AI is the next multiplier — so I'm going deep on machine learning, NLP and computer vision, with the same philosophy as always: learn it by building something with it.

The first target is already chosen — bringing intelligent, adaptive experiences into BarhtaFit. Software that learns from its users.

core focus

Machine Learning

Foundations first — supervised learning, evaluation, and real datasets over hype.

exploring

Natural Language Processing

Language as an interface — text intelligence inside everyday software.

exploring

Computer Vision

From pixels to predictions — the layer behind smarter fitness experiences.

destination

AI-Powered Products

The endgame: shipping models as features people actually use.

lab — experiment v0.3
converging
$ python train.py --model fitness-vision
epoch 21/50 — loss: 0.187 — acc: 0.884
epoch 22/50 — loss: 0.163 — acc: 0.897
epoch 23/50 — loss: 0.141 — acc: 0.905
epoch 24/50 — loss: 0.124 — acc: 0.912
validation loss-33.7% this run
“I don't want to just use AI APIs. I want to build products that learn.”
06 — open to internships · freelance · collabs

Have an idea
worth building?

I'm interested in building useful products, working on challenging software, and collaborating on interesting ideas. If it ships and matters — I want to hear about it.

techmarwan70@gmail.com
Say Hello
GitHub@marwankhanLinkedInin/marwankhanEmailtechmarwan70@gmail.com