OMNITAB / SELECTED SYSTEMS

Systems,
not slides.

Built across product, software, AI, automation, data, and operations — from the customer interface to the policy boundary behind it.

Inspect the work mohamed.moussa@omnitab.tech

BUILD GRAPHEND → END

PRODUCTWhat people use

SYSTEMWhat makes it work

CONTROLWhat keeps it safe

OPSWhat keeps it running

DESIGN → BUILD → GOVERN → OPERATE

ProductMulti-surface SaaS, mobile, dashboards, operator tools

SoftwareNext.js, React, TypeScript, Flutter, APIs, databases

AIAgents, governed assistants, deterministic + LLM hybrid systems

Automationn8n, Apps Script, webhooks, queues, workflow orchestration

DataOperational models, analytics, reporting, validation, audit trails

OperationsSafety gates, approvals, runbooks, health checks, deployment discipline

SELECTED SYSTEMS

Evidence lives in the architecture.

Different industries, different interfaces, same standard: clear boundaries, useful automation, and production-aware engineering.

01SHIPPED / MULTI-SURFACE

OmniTab Business Platform

A multi-tenant restaurant operating system, built as a set of focused products rather than one oversized app.

Customer ordering, point of sale, onboarding, operator tooling, platform administration, the public surface, a shared PocketBase API, and an AI proxy were designed as one connected product family. Production surfaces were deployed independently and health-verified.

ReactTypeScriptPocketBaseCloudflare Pages

ARCHITECTURE VIEW01/06

Customer

POS

Onboarding

Ops

Admin

API

Separate surfaces. Shared operating model. One product system.

02PRODUCTION / GOVERNED

Recruiting Engine + Operations Dashboard

A white-label, multi-campaign recruiting system with an operator dashboard designed around safe execution boundaries.

The dashboard exposes campaign, prospect, signal, account, channel, worker, hold, suppression, metric, and activity views through server-only gateways. Consequential controls use confirmations, idempotency, audit trails, allowlists, and an emergency-stop path instead of browser-direct provider access.

Next.jsReactTypeScriptVitestVercel

ARCHITECTURE VIEW02/06

Operator

Gateway

Policy

Workers

Providers

Audit

Built to operate real workflows without handing the browser the keys.

03INTERNAL OS / AI GOVERNANCE

Command Center + Business Brain

An internal Business OS paired with a durable knowledge layer for agent-assisted operations.

Command Center spans sales, growth, projects, recruitment, inbox, analytics, approvals, audit, agent runs, and document workflows. Business Brain governs durable knowledge across Directives → Orchestration → Execution while preserving ownership boundaries between CRM, operational state, files, automation, knowledge, and code.

Next.jsPostgresDrizzleR2Cloudflare Access

ARCHITECTURE VIEW03/06

Directives

Brain

Agents

Approval

Execution

Audit

Agents can propose. Consequential side effects cross an explicit approval boundary.

04FULL-STACK / HEALTH TECH

MediMe

A multi-platform medical management system spanning mobile, APIs, file infrastructure, provider portals, and AI-assisted triage.

The system combines a Flutter application, Node/Express/Postgres backend, FastAPI file service, provider/admin web surfaces, R2 storage, FCM, JWT/refresh flows, RBAC, audit logging, encryption, and lockout controls. DoctorBot uses an LLM for symptom extraction while differential and specialty routing remain deterministic.

FlutterNode.jsPostgreSQLFastAPILangGraph

ARCHITECTURE VIEW04/06

Patient

API

Records

Triage

Provider

Files

AI is used where it helps; clinical routing logic stays deterministic.

05PRODUCT / IN DEVELOPMENT

Sidus

An education and assessment platform built around Cambridge-style science learning rather than generic quiz mechanics.

The product direction covers student and teacher workflows, assessment modes, analytics, governed content, AI-assisted question generation, marking, explanations, and tutoring. The architecture keeps the learning corpus and assessment workflow central instead of treating AI as the product itself.

AssessmentAnalyticsAI TutoringContent Governance

ARCHITECTURE VIEW05/06

Content

Teacher

Assessment

Student

Analytics

Domain workflow first; AI supports the learning loop.

06AUTOMATION / DATA PIPELINE

Storm-Restoration Prospecting Engine

A low-cost operator-controlled prospecting system for discovery, enrichment, prioritization, and calling workflow management.

Google Sheets and Apps Script form the operator surface while n8n coordinates company discovery, NOAA/NCEI storm context, enrichment, queue prioritization, run history, and provider spend controls. The design favors free-first providers, lazy enrichment, preserved state, replaceable adapters, and deterministic logic with AI as optional augmentation.

Google SheetsApps Scriptn8nNOAA/NCEITypeScript

ARCHITECTURE VIEW06/06

States

Storm data

Discovery

Enrichment

Queue

Runs

Cost-aware automation with operator visibility at every stage.

AUTOMATION + DATA SYSTEMS

Small systems matter when they remove daily friction.

Lead verification and quality trackers

Operational dashboards and reporting

n8n + CRM integration workflows

AI-assisted internal tools

Data validation, routing, and audit trails

ENGINEERING PRINCIPLES

The part behind the demo.

Good interfaces are visible. Good boundaries usually are not. Both are designed.

Safe by default

Credentials stay server-side. High-consequence actions cross explicit policy and approval boundaries.

Deterministic where it matters

AI augments extraction, drafting, and reasoning without becoming a single point of failure.

Operators need visibility

Health, holds, history, cost, and audit evidence are treated as product features, not afterthoughts.

Integrations stay replaceable

External providers sit behind narrow interfaces so the business is not welded to one vendor.

Deployments count

A system is not finished at the screenshot. Build, security, environment, health, and rollback paths matter.

WORKING STACKSelected tools — not a dependency bingo card.

Next.jsReactTypeScriptPythonFlutterPostgreSQLPocketBaseDrizzleFastAPILangGraphn8nCloudflareVercelR2Google Apps ScriptPower BISQLGitHub

BUILDER / OPERATOR

Mohamed Moussa

Alexandria, Egypt

I build from an operations mindset: understand the real workflow, model the failure modes, then make the software remove friction without hiding control from the people running it.

Visit OmniTab main site ↗

OPERATIONAL BACKGROUND — NOT SOFTWARE PROJECT KPIs

500+monthly customer records analyzed with near-zero reporting errors

15agents managed in a team operating at 92% quality

60dwindow in which two recurring process defects were eliminated

NEXT SYSTEM

Bring the messy workflow.

We can start with the process, the data, or the ugly spreadsheet. The architecture comes after we understand what actually has to work.

Start a conversation omnitab.tech ↗

Systems,not slides.