JMJalal MatarRésumé
Jalal MatarBeirut, Lebanon · Remote-first

Full-stack developer / AI-accelerated delivery

I turn ambiguous problems into production systems.

Outcome-driven full-stack developer building across web, mobile, commerce, applied AI, and industrial automation. I use AI-accelerated workflows to learn unfamiliar domains quickly, make sound technical decisions, and move from requirements to reliable software.

View selected work Download résumé
GitHub LinkedIn Email
  1. 01Problem
  2. 02Learn
  3. 03Architect
  4. 04Build
  5. 05Validate
  6. 06Ship
Shopify App StorePublished product
Web + iOS + AndroidCross-platform delivery
AI + industrialDigital and physical systems
18-member teamTechnical leadership
20% improvementOperational efficiency

01 / Selected work

Systems with real consequences.

The strongest work is not a list of frameworks. It is a set of decisions made under constraints—then tested in commerce, operations, production, and the physical world.

01

Independent product · Commerce + applied AI

Aeyora

A Shopify-native analytics product that connects first-party customer journeys to plain-English, merchant-scoped analysis.

Live on the Shopify App StoreFounder & Full-Stack AI Developer

The operating problem

Merchants can see conflicting attribution across ad platforms and storefront analytics, but answering a focused business question still demands manual reporting or data expertise. Aeyora turns collected storefront and purchase signals into governed analysis without exposing raw query access to the merchant or the model.

What I owned

  • Product architecture across the embedded Shopify app, tracking pipeline, data layer, AI orchestration, billing controls, and production deployment.
  • Shopify OAuth, App Store distribution, web pixels, purchase webhooks, GDPR handlers, and tenant provisioning.
  • Agent-tool contracts, query safeguards, response shaping, regression tests, and the cost model behind plan quotas.

System map

Architecture simplified for clarity
  1. 01ShopifyPixel, OAuth, orders, billing
  2. 02App layerRemix BFF + tenant binding
  3. 03BigQueryMerchant-scoped event tables
  4. 04AI agentFastAPI + governed tools
  5. 05ValidateCost, SQL, quota, response checks
  6. 06AnswerMerchant-facing insight blocks
01

Tenant isolation outside the model

The authoritative tenant is resolved by the trusted application layer and bound server-side. It is not exposed as a tool argument the model can rewrite.

02

Constrained execution

Template-first, parameterized BigQuery queries apply date windows, dry-run byte estimates, row caps, timeouts, retention limits, and SELECT-only rules for eligible custom analysis.

03

Budget-aware AI

Plan allowances, analysis counts, planning budgets, execution budgets, and query cost ceilings are enforced before work reaches the expensive part of the pipeline.

04

Test the boundaries

The test suite pins tenant validation, SQL placeholders, authentication failures, sandbox import restrictions, curated runtime modules, and API error behavior.

ShopifyRemixReactPythonFastAPIBigQueryPostgreSQLPrismaLangChainSupabaseRailwayPytest

Result / status

Published on the Shopify App Store with an embedded install flow, merchant-level plans, first-party event collection, and production AI analysis controls.

Commercial source is private; product status is publicly verifiable.View on Shopify
02

Technical leadership · Community platform

Jama3etna

A cross-platform community, professional-networking, and learning ecosystem delivered across mobile and web by an 18-member team.

Production ecosystemFounder & Technical Lead, part-time

The operating problem

The product spans community identity, professional networking, learning commerce, payments, course delivery, certificates, and administration. The engineering challenge is not a single screen—it is keeping identity, permissions, and lifecycle state coherent across mobile, web, and service boundaries.

What I owned

  • Technical direction, delivery sequencing, system integration, and release accountability across an 18-member cross-functional team.
  • Cross-platform behavior across React Native mobile clients, the web experience, and Node/Next.js service layers.
  • Learning-marketplace workflows including SSO, payment state, enrolment, progress, instructor payouts, and certificate issuance.

System map

Architecture simplified for clarity
  1. 01Mobile + webReact Native and Next.js surfaces
  2. 02IdentityCommunity SSO + token exchange
  3. 03Platform APINode/Next.js business workflows
  4. 04Data rulesTenant context + row-level policy
  5. 05CommerceCheckout, payouts, webhooks
  6. 06GovernanceProgress + certificate lifecycle
01

Shared identity, bounded trust

The marketplace integration exchanges signed community identity through explicit server routes, validates claims, and separates browser, mobile, and backend trust boundaries.

02

Lifecycle-driven learning

Course publication, purchase, enrolment, lecture progress, resubmission, completion, and certificate issuance are modeled as governed workflows rather than loose UI state.

03

Platform-specific delivery

Mobile and web share the operating model while retaining dedicated authentication callbacks, content endpoints, and interaction patterns where the platforms differ.

React NativeNext.jsTypeScriptNode.jsExpressSupabasePostgreSQLStripePayPalPlaywright

Result / status

Shipped as a multi-platform product with community identity, production learning flows, commerce, and certificate governance—while making overlapping team work legible and reviewable.

Selected repositories are organisation-owned and may require access.GitHub organisation
03

Industrial automation · Production systems

Flo-Smart Beverage Solutions

Connected digital ordering, operational software, and beverage-machine control into a production system that improved operational efficiency by 20%.

Deployed operational systemsR&D Associate / Lead Project Engineer

The operating problem

A paid digital order only matters if it becomes the correct physical output, at the right machine, with a visible operational state. The system had to bridge commerce software and industrial hardware while remaining diagnosable for production teams.

What I owned

  • Operational dashboard and Square integration for receiving and routing incoming digital orders.
  • Integration across application services, IoT sensors, embedded controllers, PLC/HMI logic, and machine workflows.
  • Commissioning, production diagnostics, failure investigation, and validated cleaning-cycle operation.

System map

Architecture simplified for clarity
  1. 01SquareIncoming digital order
  2. 02OperationsRoute, validate, expose state
  3. 03Control layerPLC / HMI / embedded logic
  4. 04MachineBeverage production workflow
  5. 05SensorsState, quality, diagnostics
  6. 06FeedbackOperator status + recovery
01

Treat the physical system as the product boundary

The software flow was designed around real machine states, operator actions, and recoverable failure modes—not only whether an API returned 200.

02

Make production observable

Sensor networks and dashboard diagnostics gave operators a shared view of order state, machine behavior, and the point at which intervention was required.

03

Validate what matters operationally

Cleaning workflows were commissioned against ATP/CFU validation requirements, connecting software sequencing to measurable production hygiene outcomes.

SquareNode.jsReactIoTPLCHMIEmbedded systemsSensor networks

Result / status

A production integration layer spanning online orders and physical-machine execution, with a documented 20% operational-efficiency improvement.

Proprietary system; architecture is intentionally described at a non-sensitive level.

Selected engagements · 2022–2026

Independent full-stack & AI consulting

Project-based delivery for clients who needed one person to move from an unclear brief to a working system—and remain accountable after the first demo.

Community engineering · Local-first computer vision

KLW community ALPR

A rapid-deployment licence-plate recognition system for community safety teams, designed to work with existing camera infrastructure without turning every camera into a cloud-video feed.

Pipeline

  1. USB / RTSP / HTTP camera discovery
  2. Quality gate + reconnecting capture
  3. Vehicle and plate detection pools
  4. Asynchronous local OCR
  5. Deduplicated operational event feed

Operational controls

  • Runs locally with no cloud inference dependency
  • Persists clean camera URLs separately from credentials
  • Bounds in-memory crops and detection history
  • Uses worker queues and frame dropping to preserve real-time behavior under load
PythonFlaskOpenCVYOLOEasyOCRWebSocketsONNXRTSP

02 / Approach

AI acceleration without outsourced accountability.

AI shortens the feedback loop. It does not replace engineering judgment. I use it to explore faster, implement faster, and test more surface area—while remaining responsible for what reaches production.

AI accelerates

  • 01 Architecture exploration
  • 02 Implementation
  • 03 Documentation
  • 04 Debugging
  • 05 Test generation
  • 06 Refactoring
  • 07 Domain research
  • 08 Iteration

I remain accountable for

  • 01 Requirements
  • 02 Architecture
  • 03 Technical decisions
  • 04 Verification
  • 05 Security
  • 06 Reliability
  • 07 Maintainability
  • 08 Production outcomes

“When the domain is unfamiliar, I learn the operating model first, identify the constraints, and then design the software around reality.”

Domain first. System second. Production always.

03 / Experience

One coherent delivery story.

Independent, part-time, project-based, and full-time work are labelled clearly. The through-line is systems ownership—not a pile of concurrent job titles.

  1. 01

    Jul 2026 — Present

    Independent product

    Aeyora

    Founder & Full-Stack AI Developer

    Owned product architecture and production delivery for a Shopify-native first-party analytics and AI analysis platform, now published on the Shopify App Store.

  2. 02

    Sep 2024 — Present

    Part-time

    Jama3etna

    Founder & Technical Lead

    Led an 18-member cross-functional team delivering a community, professional-networking, and learning ecosystem across web, iOS, and Android.

  3. 03

    2022 — 2026

    Project-based

    Independent practice

    Full-Stack & AI Consultant

    Designed, integrated, validated, and shipped client systems across operations, commerce, conversational AI, and enterprise platform integrations.

  4. 04

    Aug 2022 — Nov 2025

    Full-time

    Flo-Smart Beverage Solutions

    R&D Associate / Lead Project Engineer

    Bridged software, IoT, embedded systems, and industrial control—from online order ingestion to commissioned beverage-machine workflows.

04 / About

A systems engineer who moved up the stack.

I did not jump randomly from industrial engineering to software. I built software inside operational systems, connected digital orders to physical machinery, and expanded that systems mindset into web, mobile, commerce, and applied AI.

My Electrical, Electronics & Communications Engineering background at RMIT University still shapes how I work: understand inputs and failure states, define the interface between subsystems, validate the output, and make the whole thing operable.

Beirut, LebanonArabic · NativeEnglish · FluentRemote-firstOpen to relocation across MENA
01

Web systems

Next.js, React, TypeScript, APIs

02

Mobile delivery

React Native, Expo, iOS, Android

03

Applied AI

Agents, tools, evaluation, sandboxes

04

Commerce

Shopify, Square, payments, attribution

05

Data platforms

PostgreSQL, BigQuery, Supabase, Parse

06

Physical systems

IoT, PLC/HMI, embedded integration

05 / Contact

Have a difficult system to build?

Bring me the unclear brief, operational constraint, or product idea. I’ll learn the domain, define the system, and help get it into production.

Emailjalal.matar96@outlook.comPhone+961 79 064 753
Beirut, Lebanon · Remote-first · Open to MENA relocation
Download résumé