Data analysts and BI professionals
Ready to build AI systems, not just query data.

Live cohort, online
Build, deploy and ship 11 real-world agentic AI projects.
From your first agent built by hand to multi-agent systems, MCP, evals and a real SaaS product. Every project goes live.
Next batch: 17 Nov, 2016 · Small live cohorts · Masterclasses included · Demo Day
Ready to build AI systems, not just query data.
You know models, but haven't built agentic systems yet.
Moving into AI engineering and want to ship real agents.
You read code and want to prototype AI features yourself.
Before you start
New to Python or rusty? You get a foundation ebook and prerecorded Python sessions to get up to speed before the program starts. No advanced ML or framework experience needed.
Not the right fit?
This program isn't built for that. It is for working professionals who already have some technical footing and want to move specifically into agentic AI and RAG.
01Make AI answer accurately from your own documents
02Build agents that decide, use tools and connect to real systems
03Run teams of agents with LangGraph, CrewAI and Microsoft Agent Framework
04Connect agents to anything with MCP and A2A
05Prove your agents work with tracing, evals and guardrails
06Turn an AI app into a real product people can pay for
The kinds of roles where these skills are used. We teach the skills; we do not promise a job, placement or salary.
Each stage gives you a new ability, and each one builds on the stage before it.
Weeks 1–2
Explain how LLMs work, write and defend prompts, ship your first AI app
Weeks 3–4
Make AI answer from your own documents, accurately and with sources
Weeks 5–6
Build agents that make decisions, use tools and connect to real systems
Weeks 7–8
Make several agents work together, at scale, on real data
Weeks 9–10
Keep agents sharp on long tasks, prove they work, design them for people
Weeks 11–12
Turn an app into a real product, and finish with the DeskPilot capstone
15 modules across 6 stages. Open any module to see what you learn.
How LLMs actually work: the basics everything else in this course builds on.
tools Python, an LLM API (setup covered in this module)
Write prompts that work, then defend them against real attacks. Your first app goes live here.
tools Streamlit, Streamlit Community Cloud, Git, GitHub
Why your app forgot your name, and the two engineering skills that fix it.
tools Streamlit (session state), your Module 2 app
Make AI answer from your company's documents: accurately, and with sources.
tools Python, Qdrant, an embedding model, document loaders, OCR
Watch basic RAG fail on real questions, then fix it, with proof.
tools BM25, Qdrant, a rerank model, SQL
Let the model make decisions, built by hand so you see exactly how it works.
tools Python, Pydantic, FastAPI
Rebuild your agent on a real framework, and combine RAG with agents.
tools LangGraph, LangSmith or Langfuse, a web search tool
The standard way to connect AI to any tool or data source, taught on the newest version.
tools MCP Python SDK or FastMCP, MCP Inspector, MCP Registry, Claude Desktop, OAuth, LangGraph
Split the work across specialist agents, and learn when not to.
tools CrewAI, YAML, LangGraph, your Module 8 MCP servers
Bigger systems: real databases, remote agents, failures and scale.
tools Microsoft Agent Framework, A2A, PostgreSQL on Supabase, your Module 8 MCP server
The skills behind today's best AI tools: keeping long-running agents focused, fast and affordable.
tools LangGraph, CrewAI or Microsoft Agent Framework; SKILL.md skills; prompt caching
Replace "it worked when I tried it" with real evidence, and add the safety checks any real agent needs.
tools LangSmith or Langfuse, OpenTelemetry, GitHub Actions
Design agents people understand, trust and stay in control of.
tools FastAPI streaming, a web front end with cards and approval screens
Turn a working app into a real product people can sign up for and pay for.
tools FastAPI, Docker, GitHub Actions, Terraform, Vercel, Render, Railway, Google Cloud Run, Ollama
Everything you've learned, in one guided build, presented live at Demo Day.
tools Everything from Modules 4 to 14
Every project ships with a live link, a README, a short demo video, its cost per run and a business case: who uses it, what it saves, what it costs.
Each project comes from a real business area, so you finish with a portfolio, not one app.
01 · AI security
A chatbot you attack with real injection and jailbreak attempts, then defend.
02 · HR
Answers from the company handbook, with page citations.
03 · E-commerce
Hybrid search and reranking over products and reviews.
04 · Careers
Matches your resume to a job post and drafts a cover letter.
05 · Finance
Cross-checks annual reports and flags what doesn't match.
06 · Hospitality
Stock, sales, bookings and supplier orders, exposed over MCP.
07 · Startups
A crew of agents writes a launch plan, then becomes a paid product.
08 · Lending
Five agents review loan applications, then get audit-ready.
09 · Market research
Researches a whole industry over 50+ steps without losing track.
10 · Travel
Plans a trip, explains each choice, and books only with approval.
11 · Customer supportCapstone
An AI support system for a SaaS company, presented at Demo Day.
Capstone
A SaaS company's support team is drowning in tickets. You build DeskPilot, an AI support system, in five guided milestones, then present it live at Demo Day.
Each with a live link, demo video and business case.
LangGraph, CrewAI and Microsoft Agent Framework, and when to pick each.
Connect agents to tools, data and other agents, on the latest versions.
The skills behind today's best AI tools.
Tracing, evals and guardrails, run automatically in CI.
LaunchPad and DeskPilot, presented live at Demo Day.
Yours to adapt to any real problem, including your company's.
A small live cohort, plus the full masterclass series.
Applied Agentic AI
₹10
GST included
EMI option available
Leave your details and we will tell you as soon as enrolment opens.
Yes, a foundation is needed. You get an ebook and prerecorded Python sessions to get up to speed first.
No. Each is taught from zero, at the point you need it.
Companies use different ones. Building by hand first, then in three frameworks, means you can pick up any new one quickly.
The capstone is guided, so everyone builds DeskPilot with the same data. You also learn what changes when you apply these patterns to your own company.
Yes. All masterclasses are included, and you receive a certificate on completion.
A laptop, a Python environment and access to an LLM API. Setup is covered in Module 1.

Cohorts are small and seats are limited. Book a free counselling call to find your batch. EMI options available.
Notify me