Khaled is available for hire
Hire KhaledKhaled El-Fakharany
Verified Expert in Engineering
AI Engineer and Developer
Manchester, United Kingdom
Toptal member since May 4, 2020
Bio
Khaled is a principal AI engineer and CTO with 12 years of experience. He was a senior engineer on the native rebuild of a media platform used by NBC, CNN and the BBC, cutting 4K transcoding from over 30 minutes to under 2, and led an MVP onto all seven Apple platforms in about four months. As CTO since 2020, he built a VR remote desktop that streams three 4K screens with head tracking under 10 ms. Today he ships agentic AI on the Claude Agent SDK, LangGraph and MCP.
Portfolio
Self-employed
Large Language Models (LLMs), Agentic AI, AI Agents, Claude Code...
Honestum LLC
Swift, iOS, Xcode, MacOS, Core ML, iOS App Development, SwiftUI, Apple Watch...
ORKA Solutions
CTO, Technical Leadership, Team Leadership, Leadership, Technical Hiring...
Experience
- Python - 10 years
- Swift - 9 years
- Solution Architecture - 5 years
- TypeScript - 4 years
- Artificial Intelligence (AI) - 3 years
- Agentic AI - 2 years
- Large Language Models (LLMs) - 2 years
- Claude Code - 2 years
Work Experience
Founder and AI Engineer
2025 - PRESENT
Self-employed
- Put a multi-agent assistant into daily use by a startup accelerator's two leaders two days after its production project went up, drafting their emails and memos from Airtable, Jira, Microsoft 365 and meeting notes.
- Kept every outward email and status change behind a person's approval that shows the exact message and recipients, a rule enforced by tests in the accelerator assistant.
- Built an orchestration layer over Claude Code, with native SwiftUI apps for Mac and iPhone, that started 741 of 805 AI coding sessions in 29 days.
- Cut paid AI tokens per technical brief from about 4.9 million to zero with a LangGraph quality gate on a local model, and vague acceptance criteria from 27-50% to 4-11%.
- Gave a retail store with no accountant an AI accountant on the Claude Agent SDK that keeps a double-entry ledger from Odoo events and previews every money movement.
- Converted 14% of 1,281 visitors into 183 sign-ups for a public workshop teaching multi-agent systems with LangGraph and the Claude Agent SDK.
- Designed a Postgres career database where every field is a fact with its source and history, with row-level security tested by break-in attempts and local pgvector embeddings.
- Built a demo-ready car-care booking marketplace in 36 hours: a SwiftUI iOS app and a React admin on Supabase, with PostGIS zones and double-entry ledgers under tests.
- Distilled a 27B teacher model into a small student on one 16 GB GPU, from a corpus generated at a 97% accept rate. The student honoured its output contract 98.9% of the time on unseen records.
Technologies: Large Language Models (LLMs), Agentic AI, AI Agents, Claude Code, Claude Agent SDK, Claude, Agentic Frameworks, LangGraph, LangChain, LLM Integration, Generative Artificial Intelligence (GenAI), Agentic Coding, AI-assisted Development, Retrieval-augmented Generation (RAG), Vector Search, Pgvector, Model Context Protocol (MCP), Prompt Engineering, AI Architecture, AI Governance, Workflow Automation, AI Automation, Knowledge Graphs, LLM Fine-tuning, llama.cpp, Hugging Face, PyTorch, Python, FastAPI, Pydantic, TypeScript, Node.js, React, Next.js, Supabase, PostgreSQL, PostGIS, Docker, Cloudflare, Dashboards, Data Visualization, API Development, Design Systems, Automated Testing, Playwright, Vitest, Swift, SwiftUI, Odoo, Airtable, Microsoft Graph API, Telegram Bot API, Jira, System Design, Unsloth, LLM Evaluation, Harness, GitLab CI/CD, Data Architecture, AI Modeling, Vite, B2B
Engineering Lead
2024 - 2025
Honestum LLC
- Led the team that shipped an AI-powered lifestyle app's MVP across all seven Apple platforms (iOS, iPadOS, macOS, watchOS, tvOS, visionOS and CarPlay) in about four months.
- Architected one SwiftUI codebase for all seven platforms, sharing the logic and data layer and adapting each interface to its platform through conditional compilation.
- Directed the conversion of external machine-learning models to Core ML, tuned to each platform's hardware, so the app's AI features ran on the device and kept users' data private.
- Designed a voice-first CarPlay interface with SiriKit for hands-free use while driving, and a visionOS experience in RealityKit that shows the user's data in 3D space.
- Guided the app through Apple's review on all seven platforms, including its access to Health and Photos data, and vetted every third-party library's privacy policy.
- Chose SwiftData with iCloud sync for storage, so a user's data followed them across their Apple devices.
Technologies: Swift, iOS, Xcode, MacOS, Core ML, iOS App Development, SwiftUI, Apple Watch, CloudKit, iCloud, HealthKit, PhotoKit, CarPlay, Sirikit, RealityKit, VisionOS, WatchOS, tvOS, iPadOS, AppKit, UIKit, WatchKit, Core Location, Instruments, TestFlight, App Store Connect, XCTest, Machine Learning, Data Privacy, Technical Leadership, Software Architecture, Code Review, Apple Vision Pro, Siri, Wearables
CTO and Co-founder
2022 - 2025
ORKA Solutions
- Co-founded a software house and grew it to 18 developers, delivering for 15 clients over three years and reaching break-even.
- Trained the team in stack-agnostic fundamentals so it could switch stacks for each client, across React, React Native, .NET, Odoo and Shopify.
- Delivered a full ERP for a multi-branch retail chain, with offline-first cloud sync, POS integration and CCTV integration.
- Built a factory ERP for a wood manufacturer on Odoo 18 in about four months, with custom modules from sales to manufacturing, after on-site discovery at the plant.
- Launched Egypt's first fresh-meat online store on Shopify, with full Arabic right-to-left support and ordering by custom cut.
- Delivered a print-on-demand platform across a React Native app, a React web app that works offline and a desktop app for print staff, linked by QR-code uploads.
Technologies: CTO, Technical Leadership, Team Leadership, Leadership, Technical Hiring, Training, Client Management, Proposals, Requirements Analysis, Agile, Software Architecture, Solution Architecture, Odoo, Python, PostgreSQL, QWeb, React, React Native, Next.js, JavaScript, .NET, Shopify, Shopify Liquid, Electron, Progressive Web Applications (PWAs), Enterprise Resource Planning (ERP), POS, Payment APIs, Offline Data Synchronization, Right-to-Left (RTL) Scripts, QR Codes, CCTV, eCommerce, IT Project Management, Interviewing, B2B, Business Strategy, Data Architecture
CTO
2020 - 2025
CanterVR
- Grew the engineering team from 2 to 20 developers as CTO, keeping rotating freelancers productive with 46 UML diagrams and more than 100 technical documents.
- Architected the four-part platform behind a VR remote desktop: a native Swift macOS server, a Unity VR app for iOS and Android, a React and Next.js website and a Firebase back end.
- Pioneered ARKit spatial tracking in VR at under 10 ms latency and 90 frames a second, for natural head movement.
- Sustained 60-minute VR sessions streaming three 4K screens on phones, with a thermal-management system of tiered fallbacks and dynamic quality scaling.
- Cut keyboard and mouse input to under 50 ms end to end by adding Bluetooth keyboard and mouse support, turning the VR app into a productivity tool.
- Added XR passthrough at 60 frames a second while keeping VR rendering at 90, so users could see their surroundings without taking the headset off.
- Designed a Central Operation Manager, an asynchronous queue system that runs the Unity app like an enterprise platform rather than a game engine.
- Shipped a native macOS server app with working VNC and audio in 11 weeks, after pitching the move from the PyQt desktop app to native Swift.
- Promoted from senior macOS developer to CTO within four months, in November 2020.
- Took the platform to market-ready status, with acquisition offers before its public launch.
Technologies: PyQt 5, Unity3D, Technical Hiring, Software Architecture, Cloud Architecture, Database Design, Internet Protocols, Networking, Remote Desktop, VNC, Audio Streaming, Virtual Reality (VR), ARKit, ARCore, Desktop App Development, Mobile Development, Node.js, CI/CD Pipelines, Scrum, JavaScript, Python, MacOS, CTO, R&D, APIs, Minimum Viable Product (MVP), User Experience (UX), Android, React, REST APIs, Mobile Applications, API Integration, Database Architecture, System Architecture, Cloud Infrastructure, System Integration, Agile Software Development, Back-end Development, Mobile Apps, SDKs, Solution Architecture, Full-stack Development, Back-end, Swift, Unity, C#, AppKit, Firebase, Firebase Cloud Functions, Cloud Firestore, TypeScript, Next.js, Amazon Web Services (AWS), Amazon Cognito, AWS Lambda, Amazon API Gateway, Amazon DynamoDB, RealVNC, Redux Toolkit, Material UI, Electron, Google VR, UML, Mermaid, Technical Leadership, Team Leadership, Technical Documentation, Extended Reality (XR), Augmented Reality (AR), Bluetooth, Thermal Management, Jira, Git, Xcode, TestFlight, Interviewing, Rendering
macOS Developer
2023 - 2023
Quill Notes LLC
- Cut transcription delays in a macOS note-taking app within a one-month engagement, after profiling its audio pipeline with Instruments to find the bottlenecks.
- Restructured the AVFoundation pipeline with tuned buffer sizes and parallel Grand Central Dispatch queues, so audio processing ran across multiple CPU cores.
- Moved audio processing onto background queues, so the interface stayed responsive while audio was recorded and transcribed.
- Redesigned the concurrency model with custom dispatch queues and barrier operations, keeping audio data intact under parallel processing.
- Found and fixed memory-management issues in the audio processing chain.
- Added a Swift capture tool to the pipeline that lists Core Audio input devices and records from the chosen microphone.
Technologies: Objective-C, MacOS, Audio, Swift, Xcode, AVFoundation, Audio Engines, Electron, Core Audio, Grand Central Dispatch, Multithreading, Concurrent Programming, Performance Optimization, Profiling, Instruments, Memory Management, Audio Processing, Transcription, Accelerate, Git
macOS Developer
2023 - 2023
B&S Collective Pty Ltd
- Built the macOS controller that runs a commercial photo booth from a Mac mini inside it, driving its payment terminal, printer and camera as one system.
- Cut the power use of booths left running 24/7 in shopping malls with a deep-sleep mode that wakes in milliseconds when a customer walks up, using IOKit power assertions.
- Designed a hybrid IPC layer: XPC services for the critical components, for isolation and crash safety, and command-line control for manufacturer software without XPC support.
- Split the app so SwiftUI drew the kiosk interface while AppKit handled system-level work and the hardware communication.
- Shipped software that runs photo booths at weddings, media events, conferences and mall installations worldwide.
Technologies: MacOS, SwiftUI, Swift 5, Automation, Debugging, SendGrid, APIs, User Experience (UX), REST APIs, API Integration, System Integration, Agile Software Development, SDKs, Solution Architecture, Swift, AppKit, Xcode, Instruments, Git, Payment APIs, Command-line Interface (CLI), Digital Kiosks, Hardware Integration, XPC, IOKit
Senior macOS Developer
2022 - 2023
Mainsail Holding LLC
- Kept a marine navigation app dating from the 1990s working for its commercial users over 13 months, in a codebase that mixed C, Objective-C, Swift and Python.
- Revived dormant features and fixed bugs in the legacy C code, documenting what each undocumented part did before changing it.
- Built new features in Swift by refactoring the Objective-C bridging layer, so modern code could sit beside the C core without destabilizing it.
- Bridged Python libraries into the macOS app to extend it without rewriting its core components.
- Moved the app's Objective-C Backendless code into Swift and added Core Data sync of its backend data.
- Updated dependencies and third-party libraries across the multi-decade codebase without breaking its existing navigation features.
Technologies: Objective-C, MacOS, Cocoa, CocoaPods, Xcode, GitHub, Jira, APIs, REST APIs, API Integration, System Architecture, Agile Software Development, Solution Architecture, C, Swift, Python, Core Data, Backendless, GPS, Git, Instruments, LLDB, Interface Builder, App Store Connect, Legacy Code, Refactoring, Debugging, Interoperability
Technical Consultant
2022 - 2022
Seidr, Ltd.
- Gave a defense-training startup what it needed to hire a development team: a product requirements document, a technical architecture and a phased roadmap.
- Ran discovery sessions with the founder to map military and SWAT training workflows, then specified every feature of the simulation platform in the requirements document.
- Designed the system architecture in C4 context and container diagrams with UML deployment, data and security views, including where AI would drive live training simulations.
- Planned the build in phases from MVP to the full platform, with technology recommendations sized for military-grade reliability.
- Handed the specifications to the development company the client hired, in a knowledge-transfer session.
Technologies: Simulations, System Architecture, Solution Architecture, Technical Consulting, Requirements Analysis, Product Requirements Documentation (PRD), Technical Architecture, System Design, Security Architecture, Cloud Architecture, Technical Documentation, UML, C4 Model, Product Roadmaps, Product Discovery, Knowledge Transfer, AI Strategy
Senior iOS Developer
2020 - 2020
Airmee
- Delivered offline support for a courier app in six weeks, so couriers could photograph, scan and confirm deliveries inside buildings with no network.
- Built a background sync service on BGTaskScheduler with heartbeat monitoring that kept running with the phone locked in a courier's pocket and synced when the network returned.
- Designed a chronological sync queue with idempotent operations, checks on both the device and the server, and fallbacks for duplicate or invalid data.
- Combined SQLite, Core Data and NSCache in one storage layer, each holding what it suits best: offline records, the object graph and temporary photos.
- Solved it entirely inside the iOS app with zero back-end changes, which removed server-side coordination and cut deployment risk.
Technologies: Swift, iOS, Architecture, Project Design & Management, Mobile, Source Code Review, Code Review, UIKit, AVFoundation, Mobile Architecture, Accessibility, Location Services, Requirements Analysis, Software Design, Native Mobile Apps, Mobile App Development, Software Architecture, R&D, APIs, User Experience (UX), REST APIs, Mobile Applications, API Integration, Agile Software Development, Mobile Apps, Solution Architecture, Core Data, SQLite, Offline Data Synchronization, Data Synchronization, Background Service Architecture, Push Notifications, Caching, QR Codes, Git, Instruments, Xcode, TestFlight, App Store Connect
Senior macOS Developer
2019 - 2020
Axle.ai
- Cut 4K transcoding from over 30 minutes to under 2 by rebuilding a five-year-old browser-based media platform as a native macOS app in six months, from its first commit.
- Cut search across 10,000+ media assets from seconds to milliseconds with a Core Data model built on compound indexes, SQLite FTS5 full-text search and denormalized search tables.
- Moved broadcasters including NBC, CNN, the BBC and The Times from overnight batch metadata extraction to real time, bridging exiv2 and Adobe XMP into Swift through Objective-C++.
- Wrapped FFmpeg 4.1 for Swift with VideoToolbox hardware encoding for broadcast formats such as ProRes, DNxHD and XDCAM, plus filmstrip thumbnails with hover scrubbing.
- Added object recognition with a Core ML ResNet50 model on extracted frames and Deepgram speech-to-text, saving detected objects and transcripts to Core Data for search.
- Shipped Ascribe, a transcription SDK built as a reusable framework, which launched as a Dropbox integration in May 2020 and opened a content-creator market beyond enterprise clients.
- Built a tiered Grand Central Dispatch queue architecture with thermal monitoring and resource-contention control for parallel processing of files tens of gigabytes in size.
- Promoted from macOS developer to senior developer within six months, after delivering the native rebuild.
Technologies: Xcode, Git, CocoaPods, Core ML, Core Data, SQLite, C++, Objective-C, Swift, Architecture, Video Editing, AVFoundation, Core Image, Video Streaming, Audio, Technical Leadership, Leadership, Requirements Analysis, Software Design, Agile, SQL, Software Architecture, MacOS, Desktop App Development, R&D, APIs, REST APIs, API Integration, Database Architecture, Agile Software Development, AppKit, Cocoa, FFmpeg, Grand Central Dispatch, Instruments, Multithreading, Memory Management, Performance Optimization, Computer Vision, Machine Learning, Deepgram, Speech-to-Text (STT), Transcription, Dropbox API, Alamofire, Full-text Search, Metadata, Video Processing, SDKs, VLCKit, VideoToolbox, ML Kit, Data Architecture, B2B
Mobile Developer and Innovation Lead
2017 - 2018
CatchSmart
- Led an indoor positioning system that tracked more than 1,000 workers at 82% accuracy, against 61.3% and 59% for two academic teams, with 3,000 beacons deployed in a 10-day push. Rolls-Royce Finland's CTO pursued licensing it.
- Launched the CatchPay marketplace to 20 shops and 500 users in its first week, after building its native iOS app from scratch in three months as the sole iOS developer. The platform was later acquired.
- Shipped Latvia's first ARKit app in three weeks, during ARKit's first month, reaching 250+ beta testers at a marketing event with 50 face masks and animated surface scenes.
- Scaled a factory quality app to three timber plants logging 1,000 to 2,000 inspections a day, building React Native screens with custom Swift and Kotlin native modules.
- Ended the crashes, memory leaks and infinite loops in factory scanning by replacing open-source camera and barcode libraries with custom native modules.
- Kept inspections working without a connection through offline-first sync, so records stayed complete for compliance audits in plants with poor coverage.
- Made beacon positioning reliable in metal-heavy plants, where structures carried Bluetooth signals across the floor, with custom signal filtering, zone clustering and path prediction.
- Delivered an eCommerce app that passed 10,000 local downloads, one of four production apps shipped as the company's sole iOS developer.
- Rose from junior developer to reporting directly to the CEO within six months, joining executive discussions and business trips for the parent group.
Technologies: Xcode, JavaScript, Git, CocoaPods, Mobile App Development, React Native, ARKit, iOS, Swift, Mobile, Architecture, UIKit, AVFoundation, Mobile Architecture, Bluetooth Low Energy (LE), Location Services, Leadership, Software Design, Native Mobile Apps, Agile, APIs, eCommerce, Android, React, REST APIs, Mobile Applications, API Integration, Agile Software Development, Mobile Apps, Objective-C, Core Data, Core Bluetooth, Core Location, MapKit, Google Maps SDK, Kotlin, Firebase, Stripe, Alamofire, OneSignal, TestFlight, Crashlytics, Jira, Scrum, Redux, iOS SceneKit, Augmented Reality (AR), Internet of Things (IoT), Indoor Positioning Systems, Signal Processing, Raspberry Pi, ESP32, Microservices, Image Manipulation, Data Architecture, B2B
R&D Engineer
2016 - 2017
Adam.ai
- Built a person-recognition module with OpenCV that identified meeting participants from video, for an AI meeting assistant that took notes and assigned tasks.
- Benchmarked NLP and computer-vision models for accuracy and performance, and turned the results into recommendations that informed the product's technical direction.
- Tested SDK integrations for meeting transcription and participant recognition, to choose the strongest machine-learning services for the platform.
- Researched product-market fit to separate essential features from nice-to-haves, which shaped the product roadmap.
- Designed and built the assistant's first physical body from sliced MDF, the first physical form of the AI system, and presented it to the parent company's CEO.
Technologies: Mobile, Architecture, Source Code Review, Code Review, UIKit, AVFoundation, Audio, Mobile Architecture, Software Design, Native Mobile Apps, Startups, APIs, User Experience (UX), 3D Modeling, Mobile Applications, API Integration, Agile Software Development, Mobile Apps, Artificial Intelligence (AI), Machine Learning, Computer Vision, Natural Language Processing, OpenCV, Python, C++, Swift, iOS, Xcode, Git, Benchmarking, Market Research, Product Strategy, Research, SDKs, CCTV, Prototyping, AI Modeling, Business Strategy
Co-founder and Instructor
2014 - 2016
Qafeer Makerspace
- Co-founded one of the first makerspaces in the Arab world, funded by the region's first successful crowdfunding campaign.
- Taught about 500 students in hands-on engineering courses, and built a model where course projects became work for the makerspace's prototyping services.
- Built a charity vending machine on a Raspberry Pi in three days for the Ramadan TV program Khawater, then deployed it in a busy Cairo mall to collect books, toys and clothes.
- Built a mobile app that let painters control an eight-degrees-of-freedom robotic arm.
- Redesigned the Google Cardboard virtual-reality viewer for higher build quality.
Technologies: Xcode, Git, CocoaPods, Mobile App Development, iOS, Swift, Python, C++, Architecture, Source Code Review, Code Review, Startups, R&D, APIs, 3D Modeling, API Integration, Agile Software Development, Mobile Apps, Raspberry Pi, Entrepreneurship, Crowdfunding, Curriculum Development, Team Mentoring, Virtual Reality (VR), Prototyping
Experience
Multi-agent Assistant for an Accelerator's Leaders
A multi-agent assistant that drafts the emails, memos and digests for a startup accelerator's two leaders, in daily use two days after its production project went up. They needed a working system for their first cohort before the full platform was ready. I built a React web app on Supabase and Cloudflare, backed by a Node daemon that runs the agents over their Airtable programme record, Jira, Microsoft 365 and meeting notes, and shows what needs doing today. Nothing leaves without a person's approval of the exact message and recipients, a rule the tests enforce. On the live system I found and cut 180 Postgres errors and 181 gateway conflicts an hour, and replaced full Airtable reads with per-record change stamps when syncing was heading for 4.4 GB of a 5 GB monthly allowance.
Orchestration System Behind 805 AI Coding Sessions
An orchestration system for parallel AI coding agents that started 741 of 805 Claude Code sessions in 29 days. It is my own system and still in R&D. A daemon keeps a registry of the sessions, starts and resumes them with one writer per repository, and exposes all of it as MCP tools. A second daemon serves native SwiftUI apps for Mac and iPhone, with live activities, widgets and voice. Underneath sits a knowledge graph of conclusions with full-text and local-embedding search, curated by a local model that is scored against a labelled gold set. Over 34 days the sessions it ran wrote about 189,000 lines of code and 6,000 tests, and the apps shipped 30 releases in three weeks.
Local-model Quality Gate for the Jira Backlog
A quality gate between technical design and the Jira backlog that cut paid AI tokens per brief from about 4.9 million to zero. Designs were reaching the backlog with vague acceptance criteria, and planning ran through a paid API. I built a LangGraph and Claude Agent SDK pipeline that checks every design task against the standards and the indexed codebase, in parallel by dependency depth, then pushes epics, stories and subtasks to Jira. Its successor runs on a local 27B model on my own GPU rig: retrieval supplies the facts, and every citation the model writes is checked against the passages it was given. Over 12 validated runs, vague acceptance criteria fell to 4 to 11 percent, against 27 to 50 percent for the hosted-model baseline, and an offline check against 34 archived plans raised 515 flags.
AI Operations Layer for a Retail Store
An AI operations layer for a retail store with no accountant, where money moved across cash drawers, a safe, the bank and card payments. I built all of the store's software. Around its Odoo 18 database sit a Telegram assistant on the Claude Agent SDK, a double-entry ledger that takes Odoo's events without counting any twice, and an AI accountant whose typed tools always show a preview before any money moves. It reads expense notes written in Arabic. A cycle-count planner packs about 2,015 products into 56 monthly stock-count tasks, each closed only by barcode evidence. The assistant now runs on the store's own server, moved from my laptop with a written cutover and rollback plan.
Offline-first ERP for a Multi-branch Retail Chain
An ERP for a multi-branch retail chain that had run entirely on pen and paper, delivered by the software house I co-founded and led as CTO. The system is offline-first: each branch keeps working on its own when its connection drops and syncs with the cloud when the network returns, so trading never depends on the internet. The branches ran POS systems, payment machines and CCTV from different vendors, so I designed a hardware abstraction layer that gives all of them one consistent interface. We rolled it out in phases, training each department on its own workflows, and moved the whole organization from paper to digital operations, with managers seeing every branch in real time.
Odoo 18 ERP for a Wood Manufacturer
A factory ERP on Odoo 18 for a wood manufacturer in Egypt, covering every department from the first quotation to delivery, built in about four months with 9 custom modules and 163 commits. Before writing any code I spent several days on site, observing operations, interviewing staff and mapping how departments worked together. I extended Odoo's ORM and workflows in custom Python modules for the wood-manufacturing steps standard Odoo lacks, while keeping it upgradeable. A sale runs as one flow: the quotation collects the requested products, an engineer builds each bill of materials, the costed quotation goes out, and confirmation raises the manufacturing order, invoice and delivery. Actions in one department trigger the next, which ended manual re-entry between HR, accounting, supply and production.
CatchPay iOS Marketplace, 500 Users in Week One
A native iOS marketplace that launched in Riga to 20 shops and 500 users in its first week. I built it from scratch in three months as the sole iOS developer, working with one back-end developer, after making the case for native Swift over React Native. People browsed local shops, ordered online and picked up in store, where coverage was often poor, so an offline-first layer on Core Data kept orders, pickup QR codes and shop locations on the phone. A pickup reminder fired when users reached the shop, using OneSignal silent pushes and background tasks to keep location updates alive within iOS limits. I stress-tested it with 1,000 shops of 500 products each, using lazy loading and hashmap search. The platform was later acquired by another company.
Ascribe, Professional Transcription SDK (Launched as a Dropbox Integration)
A transcription SDK for macOS that opened a content-creator market, from podcasters to YouTubers, beyond its company's enterprise media clients. I designed it with a clean public interface over a layered core, so Dropbox could integrate it without changes to the company's core codebase. Audio goes to Deepgram's speech-to-text API through a networking layer with retries, exponential backoff and chunked, resumable uploads with progress, and transcripts stay in Core Data for offline use. AVFoundation and AVAssetReader stream the audio out of MOV and MP4 video and handle WAV, MP3, M4A and AAC files without loading them whole into memory. The SDK was built as a reusable framework, and the product launched as a Dropbox integration in May 2020.
BConfident, AI Video Interview Analysis
An iOS prototype for automated video interviews that read a candidate's confidence, excitement and confusion in real time, built as co-founder and technical lead of an HR startup. Written tests let candidates rehearse polished answers, and video interviews gave reviewers no objective read of how people reacted. Companies created interview questions and reviewed the answers, and candidates recorded their responses on video. I integrated an emotion-detection SDK from an MIT scholar and built in-app data collection, working with the SDK's creator to make it reliable across demographics and lighting. The back end ran on Firebase, with Firestore, Authentication and Storage, and multilingual transcription opened it to candidates beyond English.
InstaPrint, Print-on-demand Across Phone, Web and Desktop
A print-on-demand platform that let customers order prints from any device instead of carrying documents to a shop on a USB drive. I led its technical development at the software house I co-founded, across a React Native app, a React web app that works offline as a PWA, and an Electron desktop app where print staff manage the order queue. People often started an order on a computer but had the document on paper, so the web app shows a QR code tied to that order: scanning it with a phone camera captures the page and attaches it to the order automatically. Customers upload contracts and other sensitive papers, so files travel encrypted, sit behind access controls and are deleted automatically. Payments are built into checkout.
AskDucks, Marketplace Matching International Students
A two-sided marketplace that matched prospective international students with current students abroad for paid advice, built as a Flutter app for iOS and Android. Early feedback was "your advice might not apply to me", so we pivoted from general consultations to matching on destination city, university, nationality, field of study and lifestyle. Services ran from a consultation call to full-service apartment hunting, with booking, messaging and payments in the app. I wrote the whole back end in Node.js and Express on MongoDB, with JWT sign-in, consultant profiles, appointments and articles, alongside Firebase for hosting, storage and cloud functions. The MVP was finished in early 2020, just before COVID-19 closed universities to international students.
Education
2017 - 2020
Bachelor's Degree in Computer Control and Computer Science
Riga Technical University - Riga, Latvia
2014 - 2017
Bachelor's Degree in Computer Engineering
Nile University - Cairo, Egypt
Skills
Libraries/APIs
CocoaPods, REST APIs, Node.js, Core Image, React, PyQt 5, Liquid, Google Maps SDK, Stripe, Alamofire, FFmpeg, Dropbox API, llama.cpp, PyTorch, Pydantic, API Development, Playwright, Telegram Bot API, HealthKit, PhotoKit, Core Audio, OpenCV
Tools
Xcode, Git, Claude Code, Figma, Amazon Cognito, SendGrid, SOLIDWORKS, GitHub, Jira, Blender, Odoo, OneSignal, TestFlight, Crashlytics, Grand Central Dispatch, Instruments, Redux Toolkit, Mermaid, Claude Agent SDK, Claude, Vitest, Sirikit, Apple Vision Pro, LLDB, Interface Builder, Firebase Authentication, GitLab CI/CD
Languages
Swift, JavaScript, HTML, SQL, TypeScript, Objective-C, C++, Python, Dart, Java, Swift 5, C, Embedded C, Kotlin, C#, UML, Shopify Liquid
Frameworks
ARKit, React Native, UIKit, SwiftUI, Electron, Unity3D, Flutter, Core Data, Core ML, Cocoa, Unity, .NET, QWeb, LangGraph, Core Bluetooth, Core Location, MapKit, Redux, iOS SceneKit, AppKit, Next.js, Material UI, Agentic Frameworks, CloudKit, RealityKit, WatchKit, XCTest, Accelerate, C4 Model, Express.js, JSON Web Tokens (JWT)
Paradigms
Mobile App Design, Agile, Desktop App Development, Agile Software Development, Cross-platform, Requirements Analysis, Microservices, Database Design, Mobile Development, Scrum, Automation, Object-oriented Programming (OOP), Thermal Management, Model Context Protocol (MCP), Automated Testing, Concurrent Programming, Refactoring, Interoperability, Affective Computing, B2B
Platforms
Google VR, iOS, MacOS, Mobile, Bluetooth Low Energy (LE), Firebase, ARCore, Android, Apple Watch, Amazon Web Services (AWS), AWS Lambda, WatchOS, iCloud, Linux, Desktop, Shopify, Raspberry Pi, Docker, VisionOS, tvOS, Backendless, Harness
Storage
Database Architecture, SQLite, MongoDB, Amazon DynamoDB, PostgreSQL, Databases, Cloud Firestore, PostGIS, Data Synchronization
Other
Artificial Intelligence (AI), Algorithms, Mobile App Development, Project Design & Management, Architecture, Technical Hiring, Source Code Review, Interviewing, Code Review, Full-stack, Video Streaming, Audio, Mobile Architecture, Technical Leadership, Accessibility, Location Services, Leadership, Software Design, Native Mobile Apps, Startups, Software Architecture, VNC, Technical Consulting, APIs, Minimum Viable Product (MVP), User Experience (UX), Mobile Applications, API Integration, System Architecture, System Integration, Back-end Development, Mobile Apps, Solution Architecture, Full-stack Development, Back-end, Large Language Models (LLMs), Agentic AI, Image Manipulation, Video Editing, AVFoundation, CTO, eCommerce, Cloud Infrastructure, Wearables, Software Development, Machine Learning, Networking, Computer Engineering, Robotics, Computer Vision, IT Project Management, R&D, Amazon API Gateway, ML Kit, Cloud Architecture, Internet Protocols, Remote Desktop, Audio Streaming, Virtual Reality (VR), CI/CD Pipelines, Siri, Debugging, Simulations, Rendering, 3D Modeling, SDKs, iOS App Development, Audio Engines, POS, Payment APIs, Enterprise Resource Planning (ERP), PWA, QR Codes, WiFi, Pathing, Sensor Data, AI Algorithms, Firebase Cloud Functions, Supabase, Research, Data Structures, Operating Systems, Computer Graphics, Software Engineering, Signal Processing, Linear Algebra, Statistics, Calculus, Engineering Design, Entrepreneurship, Programming, Embedded Systems, Mentorship, Augmented Reality (AR), Internet of Things (IoT), Indoor Positioning Systems, ESP32, Multithreading, Memory Management, Performance Optimization, Deepgram, Speech-to-Text (STT), Transcription, Full-text Search, Metadata, Video Processing, VLCKit, VideoToolbox, RealVNC, Team Leadership, Technical Documentation, Extended Reality (XR), Bluetooth, AI Agents, LangChain, LLM Integration, Generative Artificial Intelligence (GenAI), Agentic Coding, AI-assisted Development, Retrieval-augmented Generation (RAG), Vector Search, Pgvector, Prompt Engineering, AI Architecture, AI Governance, Workflow Automation, AI Automation, Knowledge Graphs, LLM Fine-tuning, Hugging Face, FastAPI, Cloudflare, Dashboards, Data Visualization, Design Systems, Airtable, Microsoft Graph API, System Design, Unsloth, LLM Evaluation, Training, Client Management, Proposals, Progressive Web Applications (PWAs), Offline Data Synchronization, Right-to-Left (RTL) Scripts, CCTV, CarPlay, iPadOS, App Store Connect, Data Privacy, Profiling, Audio Processing, Command-line Interface (CLI), Digital Kiosks, Hardware Integration, XPC, IOKit, Product Requirements Documentation (PRD), Technical Architecture, Security Architecture, Product Roadmaps, Product Discovery, Knowledge Transfer, AI Strategy, GPS, Legacy Code, Background Service Architecture, Push Notifications, Caching, Natural Language Processing, Benchmarking, Market Research, Product Strategy, Prototyping, Crowdfunding, Curriculum Development, Team Mentoring, Inventory Management, Double-entry Bookkeeping, Financial Statements, UPC & Barcodes, Offline Caching, Video Capture, Two-sided Marketplaces, Data Architecture, AI Modeling, Business Strategy, Vite
Collaboration That Works
How to Work with Toptal
Toptal matches you directly with global industry experts from our network in hours—not weeks or months.
1
Share your needs
Discuss your requirements and refine your scope in a call with a Toptal domain expert.
2
Choose your talent
Get a short list of expertly matched talent within 24 hours to review, interview, and choose from.
3
Start your risk-free talent trial
Work with your chosen talent on a trial basis for up to two weeks. Pay only if you decide to hire them.
Top talent is in high demand.
Start hiring