Zohdi Emara
Riyadh · Indoor navigation · Applied AI

ZohdiEmara

Senior Software Engineer & Technical Project Manager

I help people find their way through hospitals, airports and city districts. For 7+ years I've built the positioning SDKs, apps, kiosks and AI assistants behind that, and led the teams that ship them.

See the work LinkedIn Get in touch
Portrait of Zohdi Emara
You are hereRiyadh, Saudi Arabia
7+years at NEARMOTION
25,000+monthly active users
10+government & enterprise clients
8production navigation deployments
30projects in this portfolio
3app languages: EN · AR · ZH
What I do

Four things, done end to end

Most of my projects need all four at once: the positioning science, the product people touch, the AI on top, and a delivery plan the client can sign off.

Indoor positioning & navigation

BLE beacon positioning, floor detection and multi-floor routing that work where GPS can't.

  • iBeacon · Kontakt
  • Multilateration · geofencing
  • Mapbox · Turf · GeoJSON

AI & LLM products

Assistants and recommenders that answer real questions for travellers, visitors and residents.

  • Claude API · LLM integration
  • Recommendation systems
  • ML: XGBoost · LightGBM · LSTM

Mobile, web & kiosk

Native iOS and Android apps, web maps and full-screen kiosks built on one shared SDK.

  • Swift · SwiftUI · Kotlin
  • Next.js · React · TypeScript
  • Firebase · AWS · GCP

Technical delivery

Roadmaps, OKRs, teams of up to 5 engineers, and UAT and handover that clients can own.

  • Agile · Scrum · Kanban
  • 10+ concurrent projects
  • <3% post-release defects
Case studies

Selected work, in detail

The seven projects that best show the range: hospitals, airports, a financial district, tourism and real estate. Each shows the problem, what I did and the result.

Independent · Legal AI2026Solo build

Nizam · نظام

Arabic-first legal assistant · live and open to try

Ask Saudi regulations in Arabic or English and get an answer cited to the exact article. It covers 335 articles across the Personal Data Protection, Labor, Anti-Cyber Crime and E-Commerce laws.

What I built
  • Parsed the official Bureau of Experts text into 335 articles, keeping only current (amended) wording
  • Hybrid retrieval: Arabic-normalized BM25 plus multilingual bge-m3 embeddings, merged with weighted RRF
  • Cross-lingual query translation so English questions match Arabic legal wording
  • Streamed, cited answers with an evidence meter, grounded refusals, a law library and a bilingual RTL/LTR interface
Result
100% Hit@5 on a 60-question Arabic/English evaluation, 100% citation accuracy, and every out-of-scope question declined. The metrics are published in the app.
100%of questions found the right article in the top 5
JavaScriptPythonCloudflare Workers AILlama 3.3 70Bbge-m3BM25RAGSSE
Try Nizam live
Healthcare2025 – 2026Lead

KFMC Smart Navigation

King Fahad Medical City, Riyadh

One of the largest medical cities in the region: dozens of buildings, bridges and reception points. Patients and visitors needed a way to find the right clinic without asking at a desk.

What I did
  • Led system design and delivery of an iOS app with a custom routing engine
  • Building-to-building routes with an indoor / outdoor choice
  • Full-screen portrait kiosk in English and Arabic with search, route planning and live status
  • UX overhaul across explore, search, routing, analytics and map style
Result
Wayfinding time down roughly 35%, measured through help-desk requests and navigation session analytics.
~35%less time spent finding the way
SwiftNearMotion SDKBLE beaconsMapboxJavaScriptPythonAWS
Smart city2026Architect

KAFD Navigator

King Abdullah Financial District, Riyadh

A district of towers where the hard part is vertical: knowing which tower and which floor you're on, then getting you across open plazas to the next one.

What I did
  • Placed users on the right tower and floor from the venue's Bluetooth beacons
  • GPS fallback outdoors, with off-route detection and rerouting
  • 2D/3D map, floor switcher, live positioning chip, turn-by-turn guidance
  • Automated device test suites, QA evidence and a UAT script for the client's own team
Result
Shipped on the NearMotion SDK as a prebuilt XCFramework, with full handover documentation.
SwiftSwiftUIUIKitXCFrameworkCoreBluetoothGPSOpenStreetMap
Aviation · AI2026Lead

KKIA Smart Assistant

King Khalid International Airport

A chat-first travel assistant for airport visitors. It knows your flight, answers questions, and walks you to the gate.

What I did
  • Scoped use cases and channels, then designed the solution
  • Built live flight context, conversational Q&A, location discovery and wayfinding
  • Voice input, traveler profile, journey modes and animated route previews
  • Responsive desktop, tablet and mobile layouts, plus an embeddable /widget
Quality
Automated CI/CD deploys, Playwright end-to-end tests and axe accessibility checks.
TypeScriptNext.jsReactClaude APIPostgreSQLPlaywrightDockerVercel
Aviation2024 – 2026

Riyadh Airport Navigation Suite

King Khalid International Airport

One map, every surface: the passenger iOS app, the production web map, terminal screens and self-service kiosks.

What I did
  • Smart search with terminal and level filters, nearby places, category menus
  • Live event APIs and an AI map assistant on the production web map
  • English, Arabic and Chinese content
  • Python tooling to clean and recolor thousands of POIs
SwiftMapbox GL JSJavaScriptPythonCoreBluetooth
Tourism · AI2025Architect

AI Smart Tourism · Diriyah

Diriyah, Riyadh · Vision 2030 tourism

Visitors arrive with different interests and limited time. The app turns those into a day plan that fits.

What I did
  • Architected a recommendation engine (collaborative filtering)
  • Built the iOS trip planner: onboarding, preferences, maps and place search
  • Real-time itineraries from preferences, location and time available
Result
Visitor engagement up about 40%, measured through session depth and itinerary completion.
~40%more visitor engagement
SwiftUIKitPythonTensorFlowMapbox SearchCore DataLottie
Healthcare2024 – 2026

Dallah Hospitals

Dallah Healthcare, Riyadh

A hospital wayfinding app built for an international patient base.

What I did
  • Three languages: English, Arabic (full right-to-left layout) and Simplified Chinese
  • Language manager, RTL mirroring utilities and a translation service for places and navigation text
  • Beacon positioning patches and a premium map style
SwiftSwiftUIUIKitMapboxCoreBluetoothi18n / RTL
Real estate · AI2025

Roshn 3D Maps & Voice AI

Roshn · Ajdan community

An interactive 3D view of a whole community, plus a voice assistant residents can talk to about facility issues.

What I did
  • 3D buildings, trees and cars, with day, sunset and night lighting
  • Satellite and standard views, clickable buildings, shadows and height controls
  • Totell: AI voice facility management with apartment matching and confidence scores
JavaScriptMapbox GL 3DGeoJSONNode.jsSupabase
The platform underneath

From a beacon on the wall to an answer on screen

Every client app runs on the NearMotion Indoor Navigation SDK. I helped build its foundations as an engineer, then evolved it into a reusable XCFramework as a senior engineer. Automated CI/CD cut release cycles by about 50%.

SIGNALS

Sensing

  • BLE beacons (iBeacon, Kontakt)
  • GPS outdoors
  • Motion sensors
  • Beacon placement field app
SDK

Positioning & routing

  • Multilateration & floor detection
  • Multi-floor, indoor/outdoor routes
  • i18n, RTL, translated places
  • XCFramework · SPM · CocoaPods
SURFACES

Apps & kiosks

  • iOS & Android apps
  • Web maps (Mapbox GL JS)
  • Touch-screen kiosks
  • Indoor Map Editor
INTELLIGENCE

Cloud & AI

  • LLM assistants (Claude API)
  • Recommendation engine
  • AWS · GCP · Firebase
  • Navigation analytics
Project directory

Every project

All client work, platform work and internal tools, from 2017 to today. Filter by sector or search by client or technology.

Career

Read it like a building directory

From the top floor down: where I am now, how I got here, and where it started.

L2Current floor

Senior Software Engineer / Technical Project Manager

NEARMOTION · Riyadh
2023 – Present

I own architecture, delivery and client alignment for AI-enabled navigation platforms across government and enterprise clients.

  • Team leadership: lead cross-functional teams of up to 5 engineers across system design, sprint planning and code review.
  • Delivery: on-time delivery across 10+ concurrent projects, with roadmaps, OKRs and KPIs I own.
  • Quality: unit and UI testing quality gates keep the post-release defect rate below 3%.
  • AI in production: KKIA Smart Assistant, the AI map assistant and the Diriyah recommender.
  • Flagships: KFMC, KAFD, Riyadh Airport, Dallah and Roshn.
L1Foundations

Software Engineer

NEARMOTION · Riyadh
Nov 2018 – 2022

I built the core positioning features and the first generation of client apps.

  • Positioning: BLE beacon positioning with multilateration, floor detection and geofencing.
  • Wayfinding apps: Ministry of Education, Burjeel Medical City, Prince Sultan University, Johns Hopkins Aramco Healthcare.
  • Ithra: a large production iOS app covering events, bookings, payments and indoor maps.
  • Enterprise: attendance tracking for about 2,000 Tamkeen employees, and a Saudi Aramco asset platform that cut manual reconciliation by about 60%.
  • Cloud: 99% uptime on a Firebase and GCP stack across Android and iOS clients.
GGround floor

Web Developer

King Faisal University · Document Center
Jun 2018 – Nov 2018

Full Stack Developer

Enseyab
Jul 2017 – Sep 2017
  • Sharek: full SDLC of the university's document management system. Page loads about 30% faster, 40+ cross-browser defects fixed, zero critical security defects at go-live.
  • Research: internal studies to improve Sharek's performance and adoption.
  • ERP at Enseyab: ASP.NET, C# and PL/SQL. Critical bug backlog down about 45% in 3 months, and average fix time from 3 days to under 1.
Research

Master of Data Science

UCAM Catholic University of Murcia, Spain · March 2025. My dissertation applied machine learning to the same question I work on every day: how people move through a city.

Dissertation · Jan 2025

Big Data-Driven Urban Planning for Enhancing Smart City Development with Machine Learning: Traffic Congestion

I benchmarked eight regression models to predict jam length and find congestion hotspots. I also tested LSTM and ARIMA forecasting and K-Means clustering. Gradient-boosted trees won clearly; SVM did not suit the data.

193,559traffic records
8models benchmarked
0.983best R² (LightGBM)
Pythonpandasscikit-learnLightGBMXGBoostCatBoostTensorFlow / KerasARIMA

Model comparison · R²

Predicting jam length in km · higher is better

Axis starts at 0.85 so the gaps between models are visible.
Working with DataPython · pandas · cleaning · visualisation
Data Mining TechniquesNLP with NLTK · text mining · SQL
Statistical Data ModellingR · regression · ANOVA
Algorithms for Data ScienceKNN · Naive Bayes · clustering
Data Analytics in BusinessPower BI dashboards
Data in AI & BlockchainApplied ML · scikit-learn
Research MethodologyProposal → dissertation
BSc Computer ScienceKing Faisal University · 2018
How I work

Habits you can see in the repos

These aren't slogans. Each one comes from how my projects are actually set up.

Snapshot before I change anything

Before a big change I take a restorable backup, such as before a routing engine rewrite or a UX overhaul, so a bad day never costs the client a release.

KFMC · Dallah · tagged restore points

Test on the device, then keep the evidence

Navigation only counts if it works on a real phone in a real building. I run scripted suites for GPS, location sources and routing, and keep dated screenshots.

KAFD QA evidence · 4 automated suites

Hand over so the client owns it

Every project ends with documentation and a UAT script the client's own team can run without me.

KAFD UAT script · KFMC iOS handover

Bilingual by default

Arabic right-to-left isn't an afterthought. It's designed in from day one, and Chinese joins where visitors need it.

EN · AR · ZH across 3 venues

Built for everyone who walks in

Hospitals and airports serve everyone, so accessibility checks run in CI, not as a final audit.

Playwright + axe in the KKIA pipeline

Measure, then decide

A data science background means I back product calls with session analytics and model benchmarks, not opinions.

Navigation analytics · MSc benchmarks
Toolkit

What I build with

Every item here appears in at least one shipped project or my MSc work.

Mobile

SwiftSwiftUIUIKitRxSwiftCombineCore DataRealmObjective-CKotlinJavaMVVMClean Architecture

Location

CoreLocationCoreBluetoothiBeaconKontakt SDKMultilaterationGeofencingARKitMapbox Maps SDKMapbox GL JSGL DrawTurfGeoJSONOpenStreetMapGoogle Maps

AI & data

Claude APILLM integrationRAGHybrid search (BM25 + embeddings)Workers AIRecommendation systemsTensorFlow / Kerasscikit-learnXGBoostLightGBMCatBoostLSTMARIMANLTKpandasRPower BISQL

Web & backend

TypeScriptJavaScriptNext.jsReactNode.jsPostgreSQLSupabaseREST APIsHTML / CSSPHPASP.NETC#PL/SQL

Cloud & delivery

AWSGCPFirebaseCrashlyticsDockerVercelCI/CDSPMCocoaPodsXCFrameworkPlaywrightGit · Bitbucket

Leadership

AgileScrumKanbanOKRs & KPIsRoadmappingStakeholder managementRisk managementBudget & scopeUAT & handoverJira
Credentials

Education, affiliations, languages

Master of Data ScienceUCAM Catholic University of Murcia, SpainMarch 2025
BSc Computer ScienceKing Faisal University, Saudi ArabiaJanuary 2018
Saudi Council of EngineersLicensed member, software engineering practice. Also invited by the College of Excellence, Riyadh (2025) as one of 8 industry evaluators to modernise its CS curriculum for 500+ students.Since 2018 · 2025
Languages
ArabicNative
EnglishC2
SpanishA1

Let's find the way together

I'm open to senior engineering and technical project management roles, and to hard wayfinding and AI problems.

za_emara@hotmail.com LinkedIn
© 2026 Zohdi Emara · RiyadhMetrics are approximate, from project analytics and client reporting. Illustrations are marked.