Skip to content

Applied AI & Product R&D Engineer

I turn ambiguous problems into working products.

8+ years building 0 to 1 systems: mixed reality surgical simulators behind three patent filings, indoor navigation live in GCC venues, and SaaS running a thousand transactions a day.

See selected work
The StepBase builder showing a generated multi-stage business workflow
StepBase, a platform that generates business applications from a written description.

8+

Years building

3

Patent filings, lead on 2

11+

Publications

1,000+

Transactions a day

Qatar, UAE

Live deployments

From first prototype to production

Most of my work starts underspecified. I spend real time in discovery, build something small enough to test, then carry it into production and measure what it actually does.

  1. Discover
  2. Prototype
  3. Architect
  4. Build
  5. Deploy
  6. Measure

Healthcare research

A clinical training gap became a mixed reality prototype, then a set of tracking and vision systems, then patent-pending technology evaluated with surgeons.

Indoor navigation

A positioning problem became an architecture, then cross-platform SDKs, then a team, then live installations across Qatar and the UAE.

Business software

An operations problem became a data model, then a full-stack product, then a system handling over a thousand transactions a day.

Selected work

StepBase

Turning plain language into working business software

The problem
Internal tools platforms still expect business users to describe their own processes as databases, screens and workflow diagrams. Most teams cannot do that translation.
My role
Architecture, AI pipeline design and full-stack engineering.
What I built
A multi-tenant platform that reads a natural language description and produces a working application: forms, data models, pages, permissions and workflows. The AI does the translation, the runtime stays deterministic and auditable.
  • Generation runs as a self-correcting loop: plan, generate, configure, validate, repair.
  • Agent blocks make routing decisions at runtime inside otherwise deterministic workflows.
  • Model layer spans OpenAI, Anthropic and OpenRouter with a tiered strategy per task.

Next.js, TypeScript, Python, Vercel AI SDK, Zod, PostgreSQL

Mixed reality surgical simulation

An eight year research programme behind four surgical training systems

The problem
Surgical training depends on scarce cadaveric and clinical time. Virtual reality alone loses the tactile feedback surgeons train on, and physical models alone cannot measure performance.
My role
R&D engineer. Lead inventor on two of the three patent filings and co-lead on the third. Tracking and rendering pipelines, computer vision, simulator architecture, proposal writing and co-authorship.
What I built
Simulators that register real surgical instruments and silicone tissue against virtual anatomy, with live segmentation, optical and electromagnetic tracking, task scoring and performance metrics.

C++, VTK, OpenCV, Unity, PyTorch, YOLOv8, SAM, Meta XR SDK

3

Patent filings

11+

Publications

4

Funded programmes

  • Laparoscopy

    C++, VTK and OpenCV with optical marker tracking and instrument registration.

  • Hysteroscopy

    A Quest 3 scope trainer with task scoring and resident assessment.

  • Open surgery

    An electromagnetically tracked Unity system with remote telementoring.

  • Robotic surgery

    Soft tissue deformation, virtual cutting and instrument kinematics in OpenGL.

NavBuddy

Indoor positioning from prototype to venue rollouts

The problem
Indoor navigation has to stay accurate across a wide range of phone hardware, inside buildings where GPS does not reach.
My role
Technical co-founder. Positioning architecture, SDK development, client integrations and technical leadership of a 3-5 person team.
What I built
Native Android and iOS positioning SDKs with React Native and Flutter wrappers, augmented reality wayfinding, and the cloud infrastructure behind them.
  • Running at Place Vendôme, DIFC, Workinton and Qatar Sports Tech.
  • One positioning core shipped across four client platforms.
  • Geofence marketing integrations with Salesforce and Bloomreach.

Android, iOS, React Native, Flutter, AWS

The NavBuddy app showing an augmented reality wayfinding route inside a venue
The Thabooq point of sale screen alongside its accounting dashboard

Thabooq

A point of sale and accounting system businesses run on daily

  • Over a thousand transactions a day in production.
  • Offline-first devices with sync and conflict resolution.
  • More than 500 automated tests and zero-downtime deployments.

React, NestJS, PostgreSQL, MikroORM, Docker

The problem
Small retail and laundry businesses need a till that keeps working when the network drops, and books that survive an audit.
My role
Designed and built end-to-end, from the data model through to production operations.
What I built
A multi-tenant cloud point of sale and ERP with offline-first devices, double-entry accounting, inventory, recipe tracking and operational reporting.

Also built

Tuggle iOS app

A focus timer and to-do app on the App Store. Plan mode takes speech or text, and an LLM edits the day's task list through tool calling.

Flutter, Swift, Firebase

PairShare Web

Peer to peer file transfer over WebRTC. No accounts, no server uploads, no size limit.

SvelteKit, WebRTC

Medical education assistant Prototype

A retrieval augmented assistant over curated medical education content.

Python, LangChain

COVID-19 clinical dashboard Analytics

A live operational dashboard used by hospital leadership through the pandemic response.

React, Power BI

Medical education automation Workflow

Ingestion pipelines and automated flows for certificate generation, admissions approvals and refreshable executive reporting.

Power Query, Power Automate

Laundry POS Desktop

A standalone point of sale for retail laundry stores, running in more than ten businesses.

Java, JavaFX, SQLite

Experience

Oct 2018 to present

R&D Software Engineer, Hamad Medical Corporation

Surgical simulation research alongside the data and workflow systems Medical Education runs on.

  • Lead or co-lead inventor on three patent-pending mixed reality surgical simulation systems, and built the tracking, segmentation and rendering pipelines behind them in C++, VTK, OpenCV and Unity.
  • Trained a YOLOv8 instrument and tissue segmentation model on ground truth generated by a SAM and XMem annotation pipeline, then exported it to ONNX and ran it live in C++ inside the renderer.
  • Built a research records tool in React, NestJS and MongoDB that replaced per-person tracking documents with a single central entry point, cutting manual reporting work by around a quarter.
  • Consolidated Medical Education reporting into Power Query pipelines and executive dashboards, and automated certificate generation and admissions approvals.
  • Technical contributor and co-author across 11+ publications and four grant-funded programmes.

Aug 2020 to present

Technical Co-Founder, NavBuddy

Indoor positioning, from the first prototype to venue rollouts across the GCC.

  • Led 3-5 engineers building native Android and iOS SDKs alongside React Native and Flutter wrappers.
  • Owned positioning architecture, augmented reality wayfinding and cloud infrastructure.
  • Ran client consultations and delivered Salesforce and Bloomreach integrations for geofence marketing.

Aug 2017 to Oct 2018

Full-Stack Developer, Garajat

Early stage startup engineering across web and mobile.

  • Built an internal operations dashboard in Angular and Firebase that improved workflow efficiency by about a third.
  • Shipped cross-platform mobile apps and integrated real-time cloud data services.

Research and patents

Patent filings

  • Mixed reality surgical simulation system

    Co-lead inventor. Published by the USPTO and EPO in 2025 QA US EU

    US20250195144A1

  • Mixed reality scope maneuvering trainer

    Lead inventor. International (PCT) application filed February 2026 PCT

  • Mixed reality interface for laparoscopic simulation

    Lead inventor. US provisional filed August 2026 US

Publications

More than 11 peer reviewed papers across IEEE, Springer, MDPI and Elsevier on surgical innovation and medical simulation. Full list on request.

Education

  • MSc Computing, Qatar University

    2023, GPA 3.88 / 4.00

    Thesis on a mixed reality surgical simulator for minimally invasive surgery.

  • BSc Computer Science, Qatar University

    2018, GPA 3.80 / 4.00

Grant funded research

  • QRDI ARG Cycle 2 , 2025 to 2028

    Technical co-author, lead developer on the prototypes

  • QRDI ARG Cycle 1 , 2024 to 2027

    Technical co-author, lead developer on the prototypes

  • QNRF NPRP Cycle 11 , 2019 to 2023

    Engineering contributor

  • QNRF NPRP Cycle 05 , 2018 to 2019

    Engineering contributor

Four competitively funded national research programmes. I co-authored the two winning QRDI proposals and led prototype development on them, and worked as an engineer on the earlier QNRF programmes.

Building difficult things where software, AI and the physical world meet.

I am looking for senior roles across applied AI, product engineering and R&D, especially work that needs fast experimentation followed by real production delivery.

  • Applied AI
  • 0 to 1 product engineering
  • Rapid prototyping
  • Healthcare technology
  • Emerging technology

Based in Doha, Qatar.