Tokenization: How LLMs Turn Text Into Something They Can Process
You type a sentence and assume the model receives it as you wrote it. It does not. What actually happens to your text before the model ever sees it.
Read 11 Aug 2026
4,000+ engineers already in the roomMLOps MENA Community
Most engineers can train a model.Almost none can ship it.
Free MLOps and AI learning for engineers across the Middle East and North Africa: structured roadmaps, tool guides, live sessions and cohort courses. For companies, we also train, staff and build.
Partners
Become a partner
Host a track, sponsor a cohort, or open your platform to thousands of engineers across MENA.
Community calendar
What is coming up
Live sessions, course lessons and deadlines, straight from the community calendar.
Subscribe to the whole calendar- Sun 1 Nov · 7:00 pm
Session 3: LLM & RAG Architecture
7:00 pmCairo8:00 pmRiyadh9:00 pmDubai - Thu 5 Novall day
K8s Course - 1
- Sun 8 Nov · 7:00 pm
Midpoint Break: Quiz + Capstone Part 1 (no lecture)
7:00 pmCairo8:00 pmRiyadh9:00 pmDubai - Thu 12 Novall day
K8s Course - 2
- Sun 15 Nov · 7:00 pm
Session 4: Agentic AI Architecture
7:00 pmCairo8:00 pmRiyadh9:00 pmDubai - Thu 19 Novall day
K8s Course - 3
Our community in numbers
Our community in numbers
Built in the open across the region — every number below links to the place it comes from.
What we do
Our goal is to make production-grade machine learning skills reachable for every engineer in the Middle East and North Africa — taught by people who do this work for a living, and free for everyone.
Pick the piece you need today — none of it costs anything. We also work with companies, and that work is what pays for this.
Live sessions
Regular live sessions on production ML, taught by engineers who ship it. Recorded and published free.
Learning roadmaps
Three structured paths from zero to job-ready, built entirely on free and open-source resources.
Open courses
Cohort-based courses with live lessons, real projects, and a certificate — delivered with Zomra.
Research support
Help for researchers on reproducibility, tooling, and the engineering behind publishable work.
Internships
Routes into real teams through our partners, so students finish with production experience.
A community that answers
Thousands of engineers across MENA who review your code, unblock you, and share openings.
Learning roadmaps
All roadmaps Three paths to production ML
Built entirely on free and open-source resources — 322 curated links across the three paths.
The Academy
Go to the Academy Learn live, with engineers who ship
Courses, free live sessions and 1:1 consultations all run on the MLOps MENA Academy. One account keeps your progress and certificates together.
Courses
All coursesOpen source
Our GitHubWhat we build in the open
Every public repository under our GitHub organisation, listed straight from GitHub.
Articles
All articles Published on LinkedIn and Medium
Introduction to Large Language Models: Understanding the Foundation of Modern AI
What a large language model actually is, and why these models are called 'large' — the foundation piece, before tokenization and everything built on top of it.
Read 1 Aug 2026
The DevOps to MLOps Transition Roadmap
You already own 60–70% of the skillset. This roadmap adds only the ML layer on top — the fastest career pivot in tech, in five phases over three to five months.
Read 8 Jul 2026
Limited time
We are also on Brainsmingle
Another place to find us — our space on Brainsmingle. For a limited time you can join free this week with the code below.
The team
Who runs this community
Practitioners who build and operate ML systems for a living, teaching what they actually do at work.
Founder
Aya Nasser Salama
Senior MLOps & LLMOps Engineer with 6+ years in AI. At Unifonic she is the central MLOps/LLMOps support across AI teams — agentic products with LangGraph and MCP, LLM evaluation and observability, and 30B+ model serving on Kubernetes. She designed and teaches "Production ML Engineering" at ITI, and built Valeo's first production RAG system.
Community Director
Basem Abusaif
AI Engineer working on 3D perception and generative AI, with 5+ years building deep learning pipelines for autonomous systems. Leads the 3D perception stack at Wakeb Data and previously productionised LiDAR simulation models at Valeo. Completing an MSc in Informatics at Nile University.
Community Director
Omar Salah
Co-runs the community day to day — sessions, and keeping the programme moving.
Plus 6 leads across six owned axes
Instructor Lead · Content Lead · Research Lead · Community Growth & Partnerships Lead · Sessions & Mentorship Lead · Platform & Social Media Lead
FAQ
All answers Questions we get every week
Yes. Sessions, roadmaps, articles, guides, and community support cost nothing, and our courses are delivered with our educational partner Zomra. We are not a charity, though: alongside the community we run a services business — training, talent outsourcing, software delivery and consultation for companies — and that paid work is what pays for the free side. It never moves community content behind a paywall.
Not for the community or the roadmaps — the Basic MLOps Engineer Roadmap starts from programming fundamentals. The MLOps Practitioner course does expect basic Python and at least one model you have trained before.
If you are new to the field, start with the Basic MLOps Engineer Roadmap. If you already work in DevOps, SRE, or platform engineering, go straight to the DevOps to MLOps Transition Roadmap — you can skip most of the basics. If you are already working in MLOps, the Seniors roadmap helps you pick a specialization.
Yes. We help researchers with reproducibility, tooling, and the engineering side of publishable work. Bring your problem to the community channels.
Contact us
Talk to the community
Questions, partnerships, sponsorship, or hiring — here is how to reach us.
Partnerships & collaboration
Talk to the founder about future partnership and collaboration.
Aya Nasser Salama
Founder
Join us
Learning MLOps alone is unnecessarily hard
Thousands of engineers across MENA are already in the room — reviewing each other’s code, sharing openings, and getting unblocked. It costs nothing.