Guide · The full syllabus · Updated September 2026
DSP AI Agent Mastery Curriculum: All 16 Modules, What You Learn and What You Build
Short answer
DSP AI Agent Mastery — Zero to Master has 16 modules in five phases. Phase 0 (Modules 1–3) covers foundations, the 7-Part Job Description and tool setup; Phase 1 (4–6) vibe coding, websites and Git; Phase 2 (7–10) agents, APIs, RAG, memory and MCP; Phase 3 (11–13) testing, security and deployment; Phase 4 (14–16) multi-tenant systems, selling and AI search. A capstone earns the certificate.
How the curriculum is built
One project, sixteen modules. You build a single AI Employee — a café ordering agent — and it grows with every module: an idea in Module 1, a job description in Module 2, a website in Module 5, an agent in Module 7, memory in Module 9, tools in Module 10, a public URL in Module 13, two cafés from one deployment in Module 14. Modules 15 and 16 teach you to sell it. The capstone is your own agent, for a real use case, and it earns the certificate. Every module is taught in Urdu and English — this is the complete AI agent course in Urdu for learners worldwide.
Modules open in order, one build at a time. There are no tests and no grades: you watch, you build, you move on. The table below is the same curriculum the course page shows and the same one its Course schema carries; this page exists so the whole syllabus can be read — and cited — in one place.
Phase 0 — Zero: Foundations (Modules 1–3)
| Module | Outcome — what you can do afterwards | You build |
|---|---|---|
| M01 AI Foundations | Explain what an LLM and an agent are, what each can't do, and pick one idea worth building. | 5 ideas through the Agent Idea Filter, one chosen with a clear user and success condition. |
| M02 Prompting & Context Engineering | Write a production-grade Job Description using the 7-Part JD: Role, Goal, Audience, Tone, Steps, Rules, Examples. | Your JD v2 with three real test inputs and a note on what you changed. |
| M03 Claude · ChatGPT · Gemini | Set up Claude, Console, Claude Code, GitHub and Vercel correctly; know when to reach for each tool. | A working Claude Project and a completed setup checklist. |
Phase 1 — Builder (Modules 4–6)
| Module | Outcome — what you can do afterwards | You build |
|---|---|---|
| M04 Vibe Coding | Build software by describing it: PLAN → BUILD ONE FEATURE → TEST → COMMIT → NEXT, with Claude Code. | Your AI Employee, Part 1 running locally, on video. |
| M05 Websites | Ship a responsive multi-page site with forms from a one-page spec. | The AI Employee's four pages, desktop and mobile screenshots. |
| M06 Git & GitHub | Version-control every project; branch, merge, recover a mistake, fix the common permission errors. | A public repo with ten meaningful commits. |
Phase 2 — Agent Engineer (Modules 7–10)
| Module | Outcome — what you can do afterwards | You build |
|---|---|---|
| M07 AI Agents | Build a real agent — Job Description + Tools + Loop — that completes a multi-step task. | A demo of your AI Employee taking a full order, plus its architecture on one page. |
| M08 APIs | Connect to the Claude API and one external service; keep keys safe; handle errors and cost. | An order sent by email/Sheets from the agent, with a clean repo history. |
| M09 RAG & Memory | Give an agent knowledge and memory using the Memory Ladder; pick the right rung for the job. | Five grounded answers from a knowledge base and memory across two sessions. |
| M10 MCP | Connect an agent to real tools — Sheets, Calendar, Gmail — through the Model Context Protocol. | An MCP tool executing from the agent, result visible in the external app. |
Phase 3 — Production (Modules 11–13)
| Module | Outcome — what you can do afterwards | You build |
|---|---|---|
| M11 Testing & Observability | Test systematically with the DSP 10-Question Test Sheet; read logs instead of guessing. | The completed sheet, three fixes with before/after, and a log. |
| M12 Security | Protect keys, block prompt injection, limit what the agent can do and spend. | A 15-point checklist signed off and ten injection attempts logged. |
| M13 Deployment | Put your AI Employee on a public HTTPS URL with the backend hosted and env vars set. | The live URL — share it in the group. |
Phase 4 — Seller (Modules 14–16 + capstone)
| Module | Outcome — what you can do afterwards | You build |
|---|---|---|
| M14 Multi-Agent & Business Automation | Turn one agent into a business system: One Agent Many Clients, notifications, and when multi-agent is worth it. | Your AI Employee serving two cafés from one deployment. |
| M15 Selling AI Solutions | Run discovery, price the work, write the proposal, deliver, get paid — as a freelancer or a one-person agency. | A discovery sheet and a proposal for a real business. |
| M16 AI Search Dominance Engine | Take one target query from research to a deployed page that Google and AI assistants cite — query fan-out, SERP intelligence, competitor gap, content architecture, technical SEO, structured data, SEO Win Score — and sell it as a monthly SEO retainer. | Your first Dominance page: one query → live URL → SEO Win Score. Open from day one, no prerequisites. |
| CAP Capstone — your own agent | Build, test, secure, deploy and pitch an original agent for a real use case — not the café AI Employee. | Live URL, repo, 3-minute demo, 1-page proposal — reviewed by the DSP team. This earns the certificate. |
The frameworks the modules teach
Three of the modules are built around a named DSP framework. Each has its own page, written so you can use it before you enrol:
- M01 · The Agent Idea Filter — the five questions an idea passes before anyone builds it.
- M02 · The 7-Part Job Description — how an agent's system prompt is written at DSP.
- M09 · The Memory Ladder — which kind of memory an agent needs, and when.
What comes with the modules
- Downloads per module — templates, checklists and worksheets (the 7-Part JD template, the Agent Idea Filter worksheet, the Memory Ladder diagram, the 10-Question Test Sheet, the security checklist, the Client Acquisition Kit).
- 12 months of group support and a live debugging call every weekend, where you bring what broke.
- Three Anthropic (Claude Academy) badges — Claude 101, Claude Code 101 and Introduction to Claude Cowork — earned in Module 3 under your own name.
- The DSP Master certificate after the capstone, with a verification page that shows your live agent.
- Lifetime access — redo a module when a tool changes; re-recordings are included.
Pricing ($100 one-time / PKR 28,000), payment options and the enrolment steps are on the AI Agent Mastery page.
Frequently asked questions
How many modules does DSP AI Agent Mastery have?
Sixteen modules plus a capstone. They are grouped into five phases — Zero (foundations), Builder, Agent Engineer, Production and Seller — and open in order, one build at a time.
Do I have to do the modules in order?
Yes, with one exception. Modules open in sequence because each build extends the previous one — the café AI Employee grows from a job description in Module 2 to a multi-tenant deployment in Module 14. Module 16, the AI Search Dominance Engine, is open from day one with no prerequisites.
What do I actually build across the course?
One AI Employee — a café ordering agent — from an empty folder to a live URL: its job description (M02), its website (M05), the agent loop (M07), an email or Sheets action (M08), a knowledge base and memory (M09), an MCP tool (M10), tests and a security checklist (M11–12), the public URL (M13) and a two-café multi-tenant version (M14). Then an original agent of your own for the capstone.
How long is each module?
Long enough to finish its build. DSP does not publish lesson counts or minutes because lectures are re-recorded and re-cut as tools change; the whole course is about 30 hours of lectures, designed as roughly an hour a day for 30 days or four sessions a week for eight weeks.
What earns the certificate?
The capstone: an original agent for a real use case — not the café project — built, tested, secured, deployed and pitched, submitted as a live URL, a repo, a three-minute demo and a one-page proposal, and reviewed by the DSP team. The certificate's verification page shows the live agent.
Is the curriculum the same as the old live bootcamp?
Same café AI Employee, same frameworks, same instructor — but the live DSP Agentic Lab (closed August 2026) covered the material in seven days of Zoom classes. Mastery is the self-paced, sixteen-module version with lifetime access and twelve months of group support.
اردو خلاصہ
DSP AI Agent Mastery — Zero to Master میں 16 ماڈیول اور ایک کیپ سٹون ہے، پانچ مرحلوں میں: بنیادیں (ماڈیول 1 تا 3)، بلڈر (4 تا 6)، ایجنٹ انجینئر (7 تا 10)، پروڈکشن (11 تا 13) اور سیلر (14 تا 16)۔ ہر ماڈیول کے آخر میں آپ کچھ بناتے ہیں — جاب ڈسکرپشن، ویب سائٹ، ایجنٹ، میموری، ٹول، لائیو URL — اور وہی ایک AI ایمپلائی ہر ماڈیول کے ساتھ بڑھتا جاتا ہے۔
ماڈیول ترتیب سے کھلتے ہیں، ایک وقت میں ایک بلڈ۔ کوئی امتحان نہیں، کوئی گریڈ نہیں۔ کیپ سٹون میں آپ اپنا الگ ایجنٹ بناتے، ڈیپلائے کرتے اور پیش کرتے ہیں — اسی سے سرٹیفکیٹ ملتا ہے۔ فیس PKR 28,000 یک مشت، تاحیات رسائی۔ استاد Sardar Ghaffar ہیں، اردو اور انگریزی میں۔
Ready to build your first AI agent?
DSP AI Agent Mastery — Zero to Master: 16 modules in Urdu and English, one real AI Employee from zero to a live URL, $100 one-time, lifetime access.