Full-stack × agents
عربي

Open to roles, agent builds, and workshops

Sultan Alfaifi سلطان الفيفي

I build AI agents that check with you before they act.

I study software engineering at Umm Al-Qura in Makkah, graduating in 2027, and I’m a COOP intern at Salla. Masari, my CV builder, has about 1,600 active users. You can try most of what’s below today.

See my work Get my CV Say hi Tell me your idea

to jump anywhere

Hi, I’m Sultan.Makkah, Saudi Arabia

COOP intern at Salla · Makkah

  • Top 5% · KAUST Academy AI specialization
  • Team 5/5 · Final stage of the IBM AI Lab
  • AutoFlow · Amd FinTech Hackathon Finalist
  • WAREVA · Miyahathon 3 modeling race and final
  • 1st · Interface design challenge
  • UQU rep · Selected for the National History Lab
  • 6.5K+ followers on X & LinkedIn
  • Mustakshif · 160+ release downloads
  • Mashhad · open-source Agent Skill · 14 GitHub stars
  • AccessFinish · Nebius Global AI Hackathon entry

Try these first

Pick the one closest to your problem. Every card links to something you can try or read today, and “Inside the build” shows the call I made and how to check it.

Mustakshif application icon

Which models can
your PC actually run?

  1. Your hardware
  2. Your goals
  3. Picks with reasons
Local AI model advisor

Windows release · v0.4.0

Mustakshif

Tells you which local AI models should run well on your Windows PC, ranked for your language and what you need, with the reason behind each pick. A desktop app in Arabic and English, plus a CLI.

Creator · Product & Python engineering

  • Python
  • Windows
  • Ollama
  • CLI
Release downloads
160+
GitHub stars
11
Inside the build

The problem

Model size alone won’t tell you whether a model fits your machine or your task.

The call I made

I wrote one scoring engine for both the app and the CLI, so they give the same answer. Models come from the official Ollama library, it works offline from a cached catalog, and nothing installs until you say yes.

What you can check yourself

A published Windows installer, documented scoring rules, and public tests. The installer isn’t signed yet, so SmartScreen may warn you. The picks are estimates, and I haven’t calibrated them against benchmarks yet.

Download for WindowsSource code
The build passed.
Can a keyboard-only user book?
  1. Nemotron patch
  2. Sealed sandbox build
  3. Keyboard-only booking
  4. Database & fresh browser

Real keyboard bug · “booked” pages with the wrong day saved

15 / 15 caught

None of them counted as a fix. Every run is published, failures included.

Verified AI code repair

Nebius Global AI Hackathon entry · v0.2.0 in progress

AccessFinish

An AI patch can pass the build while a keyboard-only user still can’t book. AccessFinish runs each NVIDIA Nemotron fix in a sealed sandbox, then checks the saved booking in the database and a fresh browser before calling it fixed.

Creator · Agent engineering & evaluation

  • NVIDIA Nemotron
  • Nebius Token Factory
  • Docker
  • Node.js
  • Evaluation
Fixes verified on a fault I seeded
7 / 7
Early trials that passed in full
2 / 5
Inside the build

The problem

A green build or a confirmation page won’t tell you whether a keyboard-only user finished the task. In this app, the page could say “booked” while the database held the wrong day.

The call I made

I let the task outcome decide. In v0.1, I hashed the prompt, source, test data, and evaluator before any run and approved each proposal by its hash. The current loop retries on its own, up to 8 times, but every attempt is built and tested in sealed containers with no network, against a test copy of the app. A fix counts only when the saved booking is right.

What you can check yourself

Earlier evaluation: five frozen single-shot trials on one keyboard case. 2 of 5 passed the full procedure, and every failure is published with raw evidence. You can replay those recorded runs in the browser without any model calls. Live runs on 24 Oct, on a pinned copy of Easy!Appointments with synthetic data: 7 of 7 verified on a fault I seeded on purpose, and 0 of 7 on a real upstream bug across 43 attempts.

Read the evidenceSource code

مَشهد

  1. Brief
  2. Scenes
  3. Sound
  4. Render
  5. Review

No music unless you ask for it, and Arabic letters stay connected.

Open-source Agent Skill for motion design

Open source · v1.2.0 · Codex & Claude Code

Mashhad

Write a short brief, and Mashhad walks Codex or Claude Code through scenes, sound, and render, then a check of the frames and the rendered file before the final cut. Arabic text keeps its direction and its connected letters. An open-source Agent Skill.

Creator · Agent Skill design & motion direction

  • Agent Skills
  • Remotion
  • Python
  • Arabic Typography
  • ElevenLabs
Inside the build

The problem

Coding agents can render video, but left alone they skip the brief, break Arabic letters apart, and add music nobody asked for.

The call I made

I kept it to one portable skill: a focused brief first, one main renderer picked per job, no music unless you ask for it, and a check of the frames and the rendered file before the final render.

What you can check yourself

Public source and an Arabic demo with sound effects only that you can rebuild from its source. Claude Code support follows the documented Agent Skills format. I haven’t tested it in an interactive Claude Code session yet.

Watch the demoSource code
Ablution water in,
irrigation out only when it’s safe.
  1. Collect
  2. Treat
  3. Check readings
  4. Irrigate

Simulated sensors

  • Sensor 1
  • Sensor 2
  • Sensor 3
Irrigation held

Each decision is logged with its reason.

Ablution water reuse for irrigation

Miyahathon 3 · Through to the modeling race and final

WAREVA

WAREVA reuses ablution and sink water from mosques and large complexes to irrigate green areas, and shuts the valve whenever a quality reading fails. I lead a team of six and build the software. The hardware comes next.

Team lead · Software & AI

  • Three.js
  • Python
  • SQLite
  • Anomaly Detection
People on the team I lead
6
Design capacity of the plant
1,000 L/h
Inside the build

The problem

Reused water only helps if it’s safe. The maintenance technician running the plant needs to see the water’s quality, and irrigation has to stop on its own when a reading goes wrong.

The call I made

Irrigation is held whenever a reading fails, and every allow or hold decision is logged with its reason. The 3D plant runs in the browser, backed by a local Python and SQLite service with logs and exports.

What you can check yourself

A live 3D demo, a demo video, and public source at v0.3.0. The anomaly detector has only been tested on public datasets so far. The project started as MAWRID (مورد), and the team took the name WAREVA after a merger in October 2026.

Try the 3D plantSource code
TextVoice
Same draft either way

Your rule, read back as one sentence.

Draft · Off
Nothing turns on
until you approve it.

Hackathon prototype · Money moves are simulated

Arabic financial automation prototype

Hackathon Amd finalist · Money moves simulated

AutoFlow

Tell AutoFlow in Arabic, by text or voice, what should happen with your money, like a payday routine. It reads the rule back as one plain sentence, and nothing it drafts turns on until you approve it. Money moves are simulated.

Creator · Product, interface & AI integration

  • React
  • JavaScript
  • OpenAI APIs
  • WebRTC
Inside the build

The problem

A money instruction has to become a rule you can read, check, and approve yourself before anything runs.

The call I made

I made text, voice, and the manual editor all write to one validated workflow schema. Anything the AI drafts is saved inactive and marked “needs review”.

What you can check yourself

A live demo, public source, and tests for the schema, the API, and voice. Money moves are simulated. There’s no login, no database, and no real bank connection.

Try it liveSource code
Write your CV here,
download a clean PDF.
Your detailsExperience and skills
Live preview
Vector PDF export

No account, and your details stay on your device.

Private CV builder

Live on GitHub Pages since Feb 2026

Masari

Write your CV in the browser, in Arabic or English, and download a clean PDF. No account, and your details stay on your device.

Creator · Product & front end

  • HTML5
  • CSS3
  • JavaScript
  • pdfmake
  • localStorage
Active users
About 1,600
CVs downloaded
1,250+
Inside the build

The problem

You want to see your CV take shape as you type and leave with a usable PDF, without making an account or uploading your details.

The call I made

I kept editing and storage in the browser, wired the form to a live preview, and generate vector PDFs on your device with pdfmake.

What you can check yourself

The app and its source are public, so you can try the editor, the live preview, and PDF export right now. The usage numbers are my own count and aren’t audited.

Try it liveSource code

Smaller builds

Smaller or less finished work: a multi-agent idea engine, an Arabic Alexa skill, my graduation project, KAUST coursework, and an Agent Skills repo with one skill so far. Each card says how far it got.

Engineering alpha · v0.4.1 · Source-available

RAJIH Engine

Multi-agent decision workflow

Give it a hackathon challenge brief. Several AI agents draft ideas on their own, then each idea goes through a structured critique and an evidence check. Close or weakly supported calls stop for a person, and every run is saved, so you can see why an idea won.

  • Python
  • Multi-Agent Workflows
  • Structured Outputs
  • CLI
Read the architectureSource code

Approved Alexa skill · Public source

DANNA

Arabic AI voice assistant

An Arabic voice assistant on Alexa, built for the Saudi dialect. It keeps track of what you said earlier and answers in context.

  • Python
  • Alexa Skills Kit
  • AWS Lambda
  • LLM Integration
Source code

Graduation project · UQU

Tabayun

Legal differences for visitors

Visiting Saudi Arabia? Tabayun compares your home country’s laws with Saudi regulations and points out the differences you should know. It’s my graduation project, supervised by Dr. Ahmed Alharthi. It isn’t public yet, and it doesn’t replace legal advice.

Coursework · KAUST Academy

KAUST AI projects

Deep learning coursework

Four PyTorch computer-vision projects from the KAUST Academy AI specialization. One is a CNN I built from scratch for potato leaf disease, with 98.61% best validation accuracy in a single run.

  • Machine Learning
  • Deep Learning
  • Computer Vision
  • Neural Networks
Source code

Open source · One skill so far

Sultan AI Skills

Open-source Agent Skills

A small MIT-licensed repo of Agent Skills for Codex and Claude. One skill so far: a multiple-choice question generator, with scripts that check option lengths and place the correct answer at random.

  • Agent Skills
  • Python
  • Validation
Source code

Recognition and numbers

Where I was selected or ranked, and the numbers behind my work.

Followers across X and LinkedIn
6,500+4.5K+ on X · 2K+ on LinkedIn
Masari active users
~1,600My own count, not audited
Mustakshif release downloads
160+GitHub release download count
Volunteer hours
180+Across 11 initiatives · my own record

  1. Top 5%

    KAUST Academy bilingual logo

    KAUST Academy AI specialization

    KAUST Academy

    I got in through several assessments, then did the in-person advanced stage on machine learning and deep learning in PyTorch.

  2. Team 5/5

    IBM logo

    Final stage of the IBM AI Lab

    IBM AI Experiential Learning Lab

    One of 600 learners selected for the final stage of the IBM AI Experiential Learning Lab. Our team’s multi-agent healthcare project, Afiyah, scored 5/5 on every criterion.

  3. AutoFlow

    Amad Hackathon Arabic logo

    Amd FinTech Hackathon Finalist

    Alinma Bank × Tuwaiq Academy

    I took AutoFlow to the onsite final of Hackathon Amd, run by Alinma Bank and Tuwaiq Academy.

  4. WAREVA

    Through to the modeling race and final hackathon

    Miyahathon 3 · Oct 2026

    My team, WAREVA, made it through the digital camp and the in-person camp in Rabigh. Miyahathon is part of the innovation program led by the Saudi Water Authority, and its final runs in Jeddah in December 2026.

  5. 1st

    Umm Al-Qura University bilingual logo

    First Place, Program Interface Design Challenge

    UQU Computer Club · Oct 2025

    My interface design came first in the club’s challenge, which ran during Skills Week.

  6. UQU rep

    Selected for the National History Lab

    King Abdulaziz Foundation (Darah) · Feb 2025

    The King Abdulaziz Foundation for Research and Archives (Darah) selected me to represent Umm Al-Qura University at the national lab.

Who you’d be working with

Portrait of Sultan Alfaifi wearing traditional Saudi attire

I’m Sultan, a software engineering student at Umm Al-Qura University in Makkah. Since I launched Masari in February 2026, I’ve put a new project on GitHub almost every month. Most of them are public, so you can read the code yourself.

I build AI agents and the tools they need, mostly in Python and JavaScript, and I make them show their work. RAJIH keeps a record of every run, so you can see why one idea beat the others.

Arabic is my first language, and I build with Arabic speakers in mind. DANNA is an Alexa voice assistant built for the Saudi dialect, and AutoFlow works fully in Arabic, right to left, voice assistant included.

Download my CV
Based in
Makkah, Saudi Arabia
Education
Software Engineering, Umm Al-Qura University (graduating 2027)
Focus
Full-stack engineering and AI agents
Languages
Arabic and English
University
Umm Al-Qura University bilingual logo

How I got here

Dates come from my CV, LinkedIn, and GitHub releases. Anything without a documented date is listed under Recognition.

  1. Education

    Started software engineering at UQU

    Umm Al-Qura University, Makkah.

  2. Recognition

    Selected for the National History Lab

    The King Abdulaziz Foundation (Darah) picked me to represent UQU nationally.

  3. Course

    Generative AI for software development

    IBM certificate on using generative AI while writing code.

  4. Recognition

    First place in interface design

    UQU Computer Club challenge, during Skills Week.

  5. Course

    Data science in Python

    University of Michigan, after its four-course Python series in November.

  6. Product

    Launched Masari

    A CV builder that keeps your data on your device.

  7. Product

    Published DANNA’s source

    My Arabic voice assistant on Alexa, approved as a skill. It remembers earlier turns in the conversation.

  8. Work

    Joined Salla

    COOP training, mostly on the back end.

  9. Product

    Released Mustakshif v0.4

    Tells you which local AI models should run on your Windows PC, and why.

  10. Product

    Released RAJIH Engine v0.4.1

    AI agents draft and critique ideas, and close calls stop for a person. Source-available.

  11. Public repo

    Released Mashhad v1.2

    An Agent Skill that walks Codex or Claude Code through making a motion video, Arabic included.

  12. Hackathon

    Entered the Nebius AI hackathon

    AccessFinish checks whether an AI fix lets a keyboard-only user finish a booking. Built on NVIDIA Nemotron.

  13. Hackathon

    Qualified for the Miyahathon final

    After the in-person camp in Rabigh, my team WAREVA moved on to the modeling race.

  14. Focus

    Reliable Arabic AI agents

    I’m testing how agents use tools and memory, when they stop to ask, and how they fail. If this is your research area, email me.

Ask my AI assistant

An AI assistant that answers only from this site’s verified data. It can open a project for you, check how my work fits a role, or draft an email to me. Console commands like “run” and “help” still work.

sultan@runtime: ~

Questions you type here go to an AI assistant (Claude via OpenRouter). They're kept for 30 days without your IP to improve answers, so don't share private details. See how it's tested, failures included

How I build agents

  1. I check what actually got saved. AccessFinish counts a fix only when the booking in the database is right.
  2. Anything with consequences waits for your yes. Mustakshif asks before it downloads a model, and AutoFlow drafts stay off until you approve them.
  3. Agents get the context they need and no more. AutoFlow’s assistant never sees credentials or card details.
  4. I write down what isn’t proven yet. I haven’t checked whether RAJIH beats other approaches, and its page says so. Mustakshif calls its own picks estimates.
  5. Arabic gets the same care as English. Mashhad keeps Arabic letters connected and in the right direction.

What I do at Salla

Salla is the platform many Saudi merchants run their online stores on. I’m there for my COOP training.

Salla

June 2026 to present

Makkah, Saudi Arabia

Software Engineer Intern, COOP

Structured COOP training. Each assignment is a REST API that I build and hand in with its tests.

  • Build REST endpoints in Laravel and PHP that read and write MySQL.
  • Validate input and return clear errors.
  • Write tests for every endpoint I build.
  • React
  • JavaScript
  • PHP
  • Laravel
  • MySQL
  • REST APIs
  • Docker

React and Docker are newer to me, and I’m learning them here.

Experience · Education · Programs · Certifications · Community

Where I’ve learned and worked

Salla, Umm Al-Qura, KAUST Academy, IBM, and the groups I’ve run workshops for.

Drag the logos, or hover over one to see what I did there.Swipe the moving logos, then tap one to see what I did there.

Beyond the projectsWorkshops, courses, skills, and what I write

What I work with

Laravel, PHP, and MySQL are what I use in my Salla COOP. My public repos are mostly Python and JavaScript. The last group is what I’m learning now.

Full-stack engineering

  • Laravel
  • PHP
  • JavaScript
  • MySQL
  • REST APIs

AI agents and LLMs

  • Python
  • AI agents
  • LLM applications
  • Structured outputs
  • Multi-agent workflows

Engineering practices

  • System design
  • API integration
  • Database design
  • Git
  • Testing & debugging

Learning now

  • React
  • Docker

Courses I’ve finished

KAUST Academy, NVIDIA, IBM, McKinsey.org, DeepLearning.AI, and the University of Michigan.

KAUST Academy

KAUST AI Academy

Topics

Machine learning · Deep learning · PyTorch

What I did

I qualified through several assessments, then trained in person on machine learning and deep learning in PyTorch. My four projects from it are on GitHub.

Result

Top 5%

NVIDIA Deep Learning Institute

Fundamentals of Deep Learning

Topics

Neural networks · Deep learning · Lab exercises

What I did

I trained neural networks in the course’s hands-on labs and passed its competency-based assessment.

Result

Completed

DeepLearning.AI

Linear Algebra for Machine Learning and Data Science

Topics

Linear algebra · Math for ML

What I did

I studied the linear algebra that machine learning is built on.

Result

Completed

University of Michigan2025

Introduction to Data Science in Python

Topics

Python · Pandas · Data analysis

What I did

I used Python and Pandas to clean and reshape data and run basic statistics on it.

Result

Completed

Verify credential
McKinsey & Company

McKinsey.org Forward Program

Topics

Problem solving · Communication · Adaptability

What I did

I spent ten weeks learning to break problems down and present the answer clearly.

Result

Completed

IBM SkillsBuild

Make Agentic AI Work for You

Topics

AI agents · Everyday tasks

What I did

I covered the basics of AI agents and where they fit into everyday tasks.

Result

Completed

Other certificates

  • Data Collection and Processing with PythonUniversity of Michigan · Nov 2025Verify
  • Python Classes and InheritanceUniversity of Michigan · Nov 2025Verify
  • Python Functions, Files, and DictionariesUniversity of Michigan · Nov 2025Verify
  • Python BasicsUniversity of Michigan · Nov 2025Verify
  • Prerequisite Courses of Artificial Intelligence: Python BasicsKAUST Academy · Dec 2025
  • Introduction to SAP Basics and NavigationKing Fahd University of Petroleum and Minerals · Oct 2025
  • Google Agile EssentialsGoogle · Jun 2025Verify
  • Introduction to Project ManagementIBM · Jun 2025
  • Using Generative AI for Software DevelopmentIBM · Feb 2025

Teaching and volunteering

I ran a three-day bootcamp on building AI agents for Artificial Intelligence Pioneers, and AI workshops for Fazzah in Madinah and for Riadiat. If you want one for your group, email me.

3Trainings I’ve led
11Volunteer initiatives
180+Volunteer hours

Trainer · Artificial Intelligence Pioneers · 3 days

AI Agents Bootcamp

I covered how agents work: the fundamentals, tools, memory, planning, and multi-agent workflows. We finished by building an agent together, step by step.

Visit their site

Facilitator · Fazzah Voluntary National Association · Madinah

AI workshop

I explained the core ideas of AI in plain language and showed a few tools people can start with.

Facilitator · Riadiat

AI for business workshop

I showed entrepreneurs where AI can fit into the day-to-day work of running a business, and how to start.

Volunteering

I’ve given 180+ hours across 11 initiatives in community work, education, and tech.

I explain AI in Arabic

Mostly on X, where 4.5K+ people follow me. I cover new models and tools, like Claude versus Codex, and what they change for developers.

I write for Arabic-speaking developers deciding which AI tool to try next. Besides Claude versus Codex, I’ve covered Claude Duet and DeepSeek, and recent posts get roughly 1K to 11K views.

6,500+Followers across X and LinkedIn
DozensOne-to-one guidance sessions
01

Explainers in plain Arabic

AI agents, tools, and coding concepts, explained in Arabic without the jargon.

02

New models and tools

New releases like DeepSeek, head-to-heads like Claude versus Codex, and what each one changes in a developer’s daily work.

03

One-to-one help

Stuck on what to learn next or how to plan a project? I’ve done dozens of one-to-one sessions on exactly that.

  • AI agents
  • Claude and Codex
  • Open models like DeepSeek
  • AI tools
  • Software engineering
  • AI news

In their words

Dr. Ahmed Alharthi supervises my graduation project at Umm Al-Qura. Heyam AbuNaseer was my trainer at the IBM AI lab.

“His talent for turning requirements into practical software solutions and working well with teammates was especially notable.”

Dr. Ahmed D. Alharthi

Assistant Professor of Software Engineering

Umm Al-Qura University

Graduation project supervisor

Umm Al-Qura University bilingual logo
  • Software engineering
  • Tabayun
  • Technical skills
  • Teamwork

His full signed letter is available on request.

“I was very impressed by his performance and would be happy to recommend him for future opportunities.”

Heyam AbuNaseer

Trainer, IBM AI Experiential Learning Lab

IBM logo
  • AI agents
  • Commitment
  • Collaboration
  • Positive attitude
Read it on LinkedIn

My CV

One page in English, built from the same verified data as this site, and readable by ATS.

Download the CV

PDF · one page · English

Who's it for? (optional)

Add the company or the role, and the CV comes ready for it.

Common roles
Have the job posting? Paste it

Sultan Alfaifi

Full-Stack Software Engineer specializing in AI Agents

Makkah, Saudi Arabia | sultalfaifi@gmail.com | sultalfaifi.com · linkedin.com/in/alfaifi-sultan · github.com/SultanAlfaifi

Summary

Software engineering student at Umm Al-Qura University (graduating 2027) and COOP software engineer intern at Salla. I build AI agents that check with you before they act, and I have put a new project on GitHub almost every month since launching Masari, a CV builder with about 1,600 active users, in February 2026.

Skills

Back end: Laravel, PHP, MySQL, REST APIs

AI agents and LLMs: Python, AI agents, LLM applications, structured outputs, multi-agent workflows

Machine learning: PyTorch, deep learning, computer vision (CNNs, U-Net segmentation, transfer learning)

Practices: system design, API integration, database design, Git, testing and debugging

Languages: Arabic (native), English

Experience

Software Engineer Intern (COOP), Salla | Jun 2026 – Present | Makkah, Saudi Arabia | salla.com

  • Build REST endpoints in Laravel and PHP that read and write MySQL, delivered as structured COOP assignments.
  • Validate input and return clear errors on every endpoint.
  • Write tests for every endpoint and hand in each assignment with them.

Projects

AccessFinish, Verified AI code repair · Nebius Global AI Hackathon entry | Oct 2026 | Evidence · GitHub

  • Built a repair loop that runs each NVIDIA Nemotron fix in a sealed, network-free sandbox and accepts it only when the booking saved in the database is right.
  • Published every run, failures included: 7 of 7 fixes verified on a fault I seeded, and 0 of 7 on a real upstream bug across 43 attempts.

Mustakshif, Local AI model advisor · Windows app and CLI · v0.4.0 | Jul 2026 | Download · GitHub

  • Built a Windows desktop app and CLI in Python that ranks local AI models for a PC's hardware and the user's goals, with the reason for each pick; 160+ release downloads.
  • Wrote one scoring engine shared by the app and the CLI; it works offline from a cached Ollama catalog and installs nothing until the user says yes.

Masari, Private CV builder · live on GitHub Pages | Feb 2026 | Live · GitHub

  • Launched a CV builder in Arabic and English with a live preview and on-device vector PDF export; about 1,600 active users and 1,250+ CVs downloaded (my own count).

RAJIH Engine, Multi-agent decision workflow · Python CLI · v0.4.1 | Sep 2026 | GitHub

  • Designed a multi-agent workflow where isolated agents draft ideas, each idea gets structured critique and an evidence check, and close or weakly supported calls stop for a person.

Teaching and community

  • Trainer, AI Agents Bootcamp, Artificial Intelligence Pioneers (3 days): fundamentals, tools, memory, planning, and multi-agent workflows, ending with a live build

Recognition

  • Top 5%, KAUST Academy AI specialization (in-person advanced stage in machine learning and deep learning with PyTorch)
  • Final stage, IBM AI Experiential Learning Lab: one of 600 learners; our multi-agent healthcare project, Afiyah, scored 5/5 on every criterion
  • Finalist, Hackathon Amd (Alinma Bank × Tuwaiq Academy) with AutoFlow
  • Miyahathon 3 (Saudi Water Authority innovation program): team through to the modeling race and final, Oct 2026

Education

Software Engineering, Umm Al-Qura University, Makkah · 2023 – expected 2027

Training

  • NVIDIA Deep Learning Institute: Fundamentals of Deep Learning
  • University of Michigan: Introduction to Data Science in Python, and a four-course Python series (2025) (verify)

Generated on 27 Oct 2026 from verified data at sultalfaifi.com

Live preview. This is the page you'll download.

Want a copy by email? (optional)

Your email is used only to send this copy, after you confirm a code. Sultan gets a note with the company, the field, and the CV, plus your email if you add it.

What brings you here?

Pick the card that sounds like you. Learning agents? Start with Mashhad, a free Agent Skill for Codex or Claude Code, or Mustakshif, a free Windows app that tells you which local model your PC can run.

Hiring an engineer

You want someone who takes a ticket and finishes it, tests included.

I build REST endpoints in Laravel that validate input, handle errors, and come with tests. Mustakshif, RAJIH, and AutoFlow ship with tests you can run.

COOP at Salla since June 2026 · Graduating 2027

Email me about a role

Building an AI product

You want an agent that serves your customers in Arabic and asks before it does anything that matters.

Send me the one task it should handle and what it must never do on its own. I’ll design the tools, memory, and approval step around that, and tell you plainly what the model can’t do yet.

AutoFlow, Hackathon Amd finalist · Masari, about 1,600 active users

Tell me about your product

Running a program or community

You need someone who can teach AI agents in Arabic to people who have never built one.

I ran a three-day agents bootcamp for Artificial Intelligence Pioneers: tools, memory, planning, then a hands-on build. I’ve also given AI workshops for Fazzah in Madinah and for Riadiat.

Artificial Intelligence Pioneers · Fazzah · Riadiat · 4.5K+ followers on X

Invite me to run a workshop

Tell me what you’re working on

Send two or three lines: what you need, who it’s for, and by when. Arabic or English, whichever is easier.

sultalfaifi@gmail.com

Based in Makkah, Saudi Arabia

esc

Tip: press / anywhere. Some keys hide a secret. sultalfaifi@gmail.com

Sultan Alfaifi

عربي

Welcome

What brings you here?

Pick one and the site arranges itself around what matters to you. You can change it anytime.