Hands-on · Cohort-based · 5 onsite days × 6 hours · 2 remote consultation days
Zero experience → a live AI product → a cloud deployment you shipped yourself. Build one guided MVP from first idea to public launch, adding a working layer every day.
Mamdouh El-Nakeeb brings 14+ years of turning ideas into products, teams, and technology companies. His experience spans AI, SaaS, marketplaces, and on-demand delivery, so every lesson comes with a builder's perspective grounded in real launches. He holds an M.Sc. in AI Engineering from Quantic and a B.Sc. in Computer & Systems Engineering from Ain Shams University.
Zeew.eu
White-label infrastructure that helps businesses launch and operate their own on-demand delivery platforms.
Stunning.so
An AI-native building platform that transforms natural-language ideas into working web applications.
Viajero
A tourism marketplace designed to bring trip discovery, comparison, and booking into one experience.
Snaps AI
Personalized storybooks powered by AI, making every child the hero of their own adventure.
CTO & Co-Founder
Crete Agency · 2017 - Present
Leading technical teams that turn ambitious ideas into scalable digital products.
Partner
Zeew · 2020 - Present
Shaping the product and technology behind a global delivery SaaS platform.
Partner
Stunning.so · 2023 - Present
Helping make software creation faster and more accessible through AI.
Co-Founder
Snaps AI · 2025 - Present
Building personalized AI experiences for children and families.
Co-Founder
Viajero · 2019
Created a marketplace that made discovering and booking trips simpler.
In 5 guided days, complete beginners build one product layer by layer until it is working, tested, and deployed, followed by 2 remote consultations for individual progress.
Your product with a working AI feature, on a real URL
A real database, authentication, and access rules that don't leak
An n8n and Telegram booking workflow upgraded into a controlled custom agent
Flutter and NestJS delivered to Azure through GitHub Actions
Debug arbitrary code or build custom backends → that’s our 20-week Builder Track, the next step for the hooked.
Prerequisites
No technical background is required. Day 0 brings complete beginners to the same starting line before the onsite program begins.
Day 0 setup checklist
Install each tool, create each account, and confirm you can sign in. We start building on Day 1 instead of spending workshop time on downloads and verification emails.
Use one email address where possible. Do not share API keys or add them to GitHub, and do not enable paid cloud resources until the guided setup asks you to.
Spec
one core loop
Prompt
small asks win
Review
read the result
Iterate
one change at a time
Publish
every single day
Vibe coding gets you to a demo — engineering gets you to production.
Day 0 happens at home (~3–4 hours): install Claude Code, Codex, VS Code, and Docker; verify your GitHub, Microsoft Azure, Supabase, n8n, Telegram, Stripe, Anthropic, and OpenAI accounts; complete four 15-minute primers; and choose a marketplace category for the shared AI Service Booking Marketplace. Completing it confirms your seat.
Every day, same rhythm:Brief + recap (30 min)Concepts + demos (90 min)Applied workshop (60 min)Breaks (30 min)Build sprint (120 min)Checkpoint + demo (30 min)
Day 1
You ship: A practical software mental model and a working project shell for the AI Service Booking Marketplace.
Open Day 1 presentationDay 2
You ship: A practical AI mental model and responsive marketplace prototype with discovery, provider comparison, availability, and sample booking.
Open Day 2 presentationDay 3
You ship: An installable Flutter build and production-ready NestJS container on Azure, connected safely to Supabase and Stripe and deployed through GitHub Actions.
Open Day 3 presentationDay 4
You ship: A consent-based Telegram agent in n8n that gathers booking requirements, contacts an approved provider chat, proposes an available slot, and confirms the user's choice.
Open Day 4 presentationDay 5
You ship: A tested Telegram appointment agent that interprets doctor and salon requests, uses approved booking tools, asks for confirmation before consequential actions, and produces an auditable result.
Open Day 5 presentationAfter the onsite build
Remote consultation 1
Review each attendee's implementation, resolve technical blockers, and agree on the safest path to a complete product.
Bring: Bring the current public link, repository, architecture map, and one prioritized blocker list.
Remote consultation 2
Review the finished marketplace, user feedback, reliability evidence, and the highest-value improvements after launch.
Bring: Bring the live product, demo evidence, user observations, and a draft 30-day improvement plan.
Day 1 in detail
Use a familiar service-booking workflow to see how software turns discovery, decisions, records, and handoffs into a scalable system.
Follow one button click through the complete software system and back to the visitor.
The whole cohort builds one shared reference product: a AI Service Booking Marketplace. You can adapt the marketplace category, brand, services, providers, and visual theme while following the same discovery-to-booking architecture and instructor-led steps.
The workshop scope stays achievable: one service category, a small provider catalog, predefined availability, one booking flow, and Stripe test mode instead of real payments or provider payouts.
Day 1
A practical software mental model and a working project shell for the AI Service Booking Marketplace.
Day 2
A practical AI mental model and responsive marketplace prototype with discovery, provider comparison, availability, and sample booking.
Day 3
An installable Flutter build and production-ready NestJS container on Azure, connected safely to Supabase and Stripe and deployed through GitHub Actions.
Day 4
A consent-based Telegram agent in n8n that gathers booking requirements, contacts an approved provider chat, proposes an available slot, and confirms the user's choice.
Day 5
A tested Telegram appointment agent that interprets doctor and salon requests, uses approved booking tools, asks for confirmation before consequential actions, and produces an auditable result.
Nobody starts over each day. Every lesson is applied immediately to the same product, and every checkpoint leaves a working recovery version before the next layer is added.
No. The program is designed for complete beginners; Day 0's prework and the four 15-minute primers bring everyone to the same starting line.
No. Day 1 teaches software through familiar workflows and a food-ordering system. Day 2 introduces AI, LLMs, RAG, agents, MCP, and model choices before you apply them to the marketplace. Every new step is demonstrated first.
No. The cohort builds one shared AI Service Booking Marketplace over five onsite days. You can adapt the category, brand, services, providers, and visual theme while everyone follows the same discovery, matching, booking, payment, dashboard, and deployment architecture.
No. Vibe coding gets you a demo; this bootcamp runs the same Spec → Prompt → Review → Iterate → Publish loop every day, so you always read, test, and own what the AI generates — the engineering habits that get you to something that actually works.
You'll learn how LLMs actually work — tokens, context, and why they hallucinate — write your own system prompt, and run a 10-case eval sheet to test and improve it. You leave able to explain exactly what your AI feature does and doesn't do.
No. Telegram bots cannot initiate private conversations with arbitrary phone numbers. The workshop uses opted-in requester and provider chats that started or approved the bot, with explicit consent before information is shared or an appointment is booked.
Bring a laptop with Claude Code, Codex, VS Code, and Docker installed, plus its charger and your phone. Before Day 1, verify your GitHub, Microsoft Azure, Supabase, n8n, Telegram, Stripe, Anthropic, and OpenAI accounts. The workshop provides AI credits.
Each onsite step is demonstrated before you apply it, instructors help with setup and troubleshooting, and every day has a working recovery checkpoint. Two later remote consultations provide an architecture clinic and a product-review roadmap.
It runs in test mode: the full, real checkout experience with zero real money moving.
Then 2 remote consultations help you improve what you launched.