Currently buildingLayerDock

AI ENGINEER

Adeen Khan

I build AI-powered products end to end, from LLM features and MCP servers to the browser extensions, APIs and billing that ship them. Currently co-founding LayerDock, where I lead the engineering.

  • MCP
  • Llama
  • Next.js
  • React
  • TypeScript
  • Supabase
  • PostgreSQL
  • Chrome MV3
  • Tailwind
  • Upstash
  • Paddle
  • Resend
  • Vercel
  • Sentry
  • Vitest
  • Node.js
  • Docker
  • Framer
  • GSAP
  • Python
  • MongoDB
ABOUT

EngineeringIntelligent Systems

I am Muhammad Adeen Khan — most people call me Adeen — an 18-year-old self-taught developer from Karachi. I learned to build production SaaS by shipping real apps rather than sitting through a bootcamp, and I am on GIAIC's AI certification track, heading into a BS in Artificial Intelligence at SZABIST. Right now I am building LayerDock, a visual feedback SaaS where I lead the engineering: the Chrome extension, the Next.js web app, the Postgres database, authentication, subscription billing, five third-party integrations, and an MCP server that gives coding agents the feedback context. Alongside it I freelance on AI-powered automation, and I would rather chase an unconventional idea than another saturated trend.

AILLMs & MCP
EXTENSIONSChrome MV3
BACKEND110+ API Routes
SECURITYRLS & Encryption

Latest Works

LayerDock

(saas)

A SaaS for visual website feedback that I co-founded and lead the engineering on: a Chrome extension that pins comments onto any element of a live site, and a Next.js dashboard where teams sort through them. Includes auth, Paddle billing, Slack, Linear, GitHub, Jira and Trello sync, and AI element labeling.

TS Files
600+
API Routes
110+
Migrations
51
Commits
550+
#next.js16#chromemv3#supabase#paddle

School OS

(fullstack)

A secure, full-scale school management platform streamlining student registration, automating fee workflows, and providing real-time financial insights through a centralized dashboard.

#next.js#typescript#postgresql#nextauth.js

AI Employee

(systems)

An autonomous agent that watches seven channels at once — Gmail, WhatsApp, LinkedIn, X, Facebook, Instagram and the filesystem — decides what deserves action, and acts through MCP servers wired to email, LinkedIn and a dockerised Odoo instance. An Obsidian vault holds its memory; every action is logged and gated behind human approval.

Watchers
7
MCP Servers
3
API Routes
24
Python Files
43
#python#mcp#groqllama3.3#odoo

Lead Intelligence

(fullstack)

A lead-generation engine that scrapes businesses off Google Maps, visits each site with Playwright to detect 20+ technical faults, identifies the site builder behind it, scores the lead by sales potential, and drafts the outreach. FastAPI backend, Next.js dashboard, SQLite store.

API Routes
25
Python Files
22
Site Checks
20+
Builders Known
9
#fastapi#next.js14#playwright#sqlite

Ocean Chat

(backend)

A Gemini-backed chat platform built for production rather than demo: persistent conversation memory in MongoDB with an in-memory fallback, multi-layer safety filtering, structured JSON output with metadata, and session stats and history behind a versioned FastAPI surface. Next.js chat client on the front.

Endpoints
17
Safety Layers
4
Storage Modes
2
#fastapi#gemini#mongodb#next.js

Voice Agent

(backend)

A voice and chat agent that answers as the business: it detects intent, pulls the customer details out of natural speech, works out which service is being asked for, and runs the whole booking lifecycle against SQLite. Speech in and out, with a local Ollama model so no call audio leaves the machine.

Python Modules
12
Response Time
<3s
Speech In / Out
STT
#python#ollama#gemma2#fastapi

DetailFlow

(fullstack)

A booking and automation platform for car detailing businesses with payments, scheduling, and WhatsApp notifications.

#next.js#typescript#prisma#stripe
CONTACT

Ready to
Collaborate?

I'm currently open to new opportunities and interesting projects. Let's build something exceptional together.