NVIDIA Inception · application live · pre-launch

Production AI at the intersection of frontier models & emerging-market scale.

Bin Abdullah LLC engineers AI-native products for the $1.5T global textile industry, the UK regulated SaaS market, and the Pakistani property economy. Launching FabricScan, TextileLens and AvatarAI on NVIDIA-native infrastructure — built on the foundation of 18+ already-shipped production apps.

$1.5T

Global textile market

18+

Live production apps

6

Operating SaaS

5+

Cloud programs backing

Launching · NVIDIA-native

Three flagship products. One GPU pipeline.

Each product is engineered on NVIDIA-native compute — TensorRT, Triton, NIM, H100 — with a clear cost-per-inference scaling story.

Launching Q3 2026

FabricScan

Real-time textile defect detection at the loom

Vision Transformer + TensorRT pipeline running on NVIDIA-class edge GPUs. Targets <100 ms per frame for live in-line inspection — replacing 2-week manual QC cycles.

Closed Beta

TextileLens / SourceIQ

Agentic AI sourcing across 10,000+ verified mills

Buyer pastes a tech-pack, the agent quotes price/MOQ/lead-time across our network in seconds — replacing weeks of WhatsApp back-and-forth on Alibaba.

Public Waitlist

AvatarAI

Multilingual talking-head video API

Self-serve avatar cloning from a single selfie. Urdu, Hindi, Arabic, English. 10–100× cheaper than Synthesia. No watermark on the free tier.

Already backed by

Validated by every major cloud platform — except NVIDIA.

We're applying to NVIDIA Inception because our roadmap is GPU-bound. Below: programs that already approved Bin Abdullah LLC.

AWS Activate

$5K credits + Bedrock + EC2

Microsoft Founders Hub

$2.5K Azure + OpenAI quota

Cloudflare for Startups

Workers + R2 + D1 + KV

Modal Labs

GPU credits for inference

PostHog

Free Cloud product analytics

Anthropic

Claude API credits

How we use NVIDIA

NVIDIA-native by design.

Every roadmap item maps to a clear CUDA / TensorRT / NIM acceleration story — not GPUs bolted on as an afterthought.

TensorRT

FabricScan factory-edge inference. Sub-100ms per frame target on Jetson-class hardware.

Triton Inference Server

Multi-model serving — concurrent FabricScan + SourceIQ at the edge with shared GPU memory.

NIM Microservices

Containerised inference for B2B customer deployments on customer-controlled NVIDIA stacks.

Modulus / NeMo

Custom LLM fine-tuning — textile domain (SourceIQ) and multilingual avatar (AvatarAI).

H100 / A10G classes

AvatarAI diffusion + audio2video pipeline. Cost-per-second model with reserved-capacity scaling path.

Inception program

We are NVIDIA-native by design — every roadmap item maps to a CUDA, TensorRT or NIM acceleration story.

Founder

Muhammad Hasnain

Founder & CEO, Bin Abdullah LLC

Solo founder operating as a fully-staffed virtual company. Has personally shipped 18+ production AI/SaaS apps from Pakistan serving UK, US and global customers. 10+ years textile industry network on the ground in Faisalabad — the Pakistan manufacturing capital.

Founder credentials earned 2026

  • Microsoft Applied Skills — Create an AI agent (lab-based, 2026)
  • Anthropic — Claude Code 101 + Claude Cowork (2026)
  • HubSpot Academy — Inbound Marketing Certified (2026)
  • Qualifi Level 5 Extended Diploma in Business Management (Ofqual-regulated UK, 2025)
  • DeepLearning.AI — Generative AI with LLMs (in progress, 2026)

Let's build

Building in public.
Hiring soon.

Investors, partners, and AI/ML engineers in Pakistan and the UK — reach out. We're raising signals, hiring our first GPU-ops engineer, and looking for design partners on FabricScan.