All work

Founder-built AI SaaS

Zellio — AI Commerce Support and Engagement Platform

I founded Zellio and led the product and engineering work from architecture through live operation, building a bilingual SaaS that connects customer conversations, store data, campaigns, and AI-assisted support workflows.

Role
Founder and Product Engineer / Technical Lead
Period
2025–present

Founder-built product system

Commerce support, connected

Simplified system map

Customer conversations

Store and order data

Zellio platform

AI-assisted workflows

Evidence-led system view; not a recreation of a customer or operations dashboard. Diagram connecting customer conversations and store data to the Zellio platform, AI workflows, campaigns, and web and mobile operation.

Context

Zellio LLC

Role

Founder and Product Engineer / Technical Lead

Product type

Founder-built AI SaaS

Core capability

TypeScript

Product context

Bringing fragmented commerce support into one product

Commerce support work is spread across messaging channels, store systems, customer records, campaigns, and manual follow-up. Zellio needed to bring those moving parts into a multi-tenant product while remaining reliable against real commerce and messaging APIs.

Attribution anchor

What I was responsible for

As founder and technical lead, I shaped the product and architecture and implemented core work across the web applications, APIs, AI workflows, integrations, mobile client, testing, infrastructure, and production operations. I led and implemented the platform; I do not present it as a one-person codebase.

Selected scope

Selected contributions

  1. 01

    Built unified support and conversation workflows across WhatsApp, Instagram, Messenger, email, Telegram, and onsite chat.

  2. 02

    Delivered Shopify, WooCommerce, Zid, and Salla connections covering synchronization, webhooks, onboarding, and order workflows.

  3. 03

    Implemented AI-assisted knowledge and agent workflows, retrieval, prompt evaluation, observability, document handling, and visual workflow automation.

  4. 04

    Built campaigns and retention workflows with delivery, retry, attribution, recipient inspection, and operational controls.

  5. 05

    Developed authentication, subscriptions, analytics, internal operations, bilingual English/Arabic experiences, and an Expo support application.

  6. 06

    Established repeatable cloud infrastructure, deployment pipelines, background processing, testing, monitoring, recovery practices, and production incident handling.

Evidence, not reconstruction

System view

A simplified view of verified responsibilities and system relationships. It does not reproduce a private or historical interface.

Founder-built product system

Commerce support, connected

A simplified view of the product boundaries I led and implemented.

Simplified system map

Customer conversations

Messaging, email, Telegram, and onsite chat entry points.

Input

Store and order data

Commerce connections, synchronization, webhooks, and lifecycle events.

Input

Zellio platform

Multi-tenant product, permissions, operations, subscriptions, and analytics.

System

AI-assisted workflows

Knowledge, retrieval, agents, evaluation, and reviewable automation.

Process

Campaigns and retention

Delivery, retries, attribution, and operational controls.

Process

Web and mobile operation

Bilingual experiences for support teams and internal operators.

Output

Verified relationships

  • Customer conversationsZellio platform
  • Store and order dataZellio platform
  • Zellio platformAI-assisted workflows
  • AI-assisted workflowsWeb and mobile operation
  • Zellio platformCampaigns and retention
Evidence-led system view; not a recreation of a customer or operations dashboard. Diagram connecting customer conversations and store data to the Zellio platform, AI workflows, campaigns, and web and mobile operation.

Engineering judgment

Selected technical decisions

Decision 01

Used a TypeScript pnpm/Turborepo structure to share product, authentication, and integration code across Express, Next.js, React, and Expo applications.

Decision 02

Combined PostgreSQL with Prisma and Drizzle, Redis-backed processing, BullMQ, Inngest, and event/webhook flows for long-running and failure-prone work.

Decision 03

Integrated multiple AI providers and workflow tooling, while managing AWS and Cloudflare infrastructure through Terraform and staged delivery practices.

Technology context

  • TypeScript
  • Next.js
  • React
  • Express
  • Expo
  • PostgreSQL
  • Prisma
  • Drizzle
  • Redis
  • BullMQ
  • Inngest
  • Mastra
  • Terraform
  • AWS
  • Cloudflare
  • Playwright
  • Vitest

Delivered capability

A live bilingual product across web and mobile, demonstrating end-to-end ownership from ambiguous product idea through architecture, integrations, testing, deployment, and continued production improvement.

AI-enabled SaaS · Commerce integrations · Product engineering

Building an AI-enabled SaaS or commerce workflow?