1
Week 1 · foundations
Setup & the agent loop
Make your first model call, get reliable structured output, and learn the one diagram the whole course hangs on.
You build: your first working model call
A hands-on course where you ship 3 real AI agents in 6 weeks.
One track for everyone · Career support included · Nothing costs extra
You write code that plans, calls tools, reads documents, and takes action — from the very first lesson.
You build an agent from scratch before touching LangGraph or CrewAI — so the frameworks never confuse you again. It's all the same loop underneath.
The core module: make an agent answer from real documents, cite the exact page, and refuse to guess. This is the hard, valuable part — and our day job.
Evaluation, guardrails, memory, human-in-the-loop, and on-prem privacy — the difference between a cool demo and something you'd put in front of a customer.
Every agent is a loop: reason → act → observe → repeat. Each module adds one capability on top — concept, live build, lab, checkpoint.
Make your first model call, get reliable structured output, and learn the one diagram the whole course hangs on.
Build a tool-using agent in ~60 lines of Python — no framework — then rebuild it in LangGraph so you see exactly what the framework hides.
Make an agent answer accurately from PDFs and scans, link every answer to the exact page and line, and correctly say "not in the document."
Coordinate specialist agents with CrewAI, add a human approval gate in LangGraph, and learn when multi-agent is powerful vs. when it's overkill.
The difference between a demo and a system you'd trust: eval harnesses, tracing, guardrails, prompt-injection defense, and on-prem privacy.
Wrap an agent as an API with a simple UI, expose your tools over MCP, and build + demo one agent that solves a real problem for you or your org.
No beginner lane, no advanced lane. You just point your agents at your own problem.
See what the frameworks hide.
Finish with agents you can demo in an interview.
Non-coders welcome. Every lab has a template.
Join the next open cohort, or bring the whole workshop in-house for your team.
Shipped agents only help if people can find them and you can talk about them.
The same six modules, compressed into a few days and run on your own documents.
Two to three days, on-site or private virtual, running inside your environment. Your team practises on real documents — no sample data, no cloud upload.
Everyone builds a working agent from scratch, engineers and non-engineers alike, so the whole room shares one vocabulary.
The core session. Point an agent at your own PDFs and scans, cite every answer to its exact page, and refuse to guess.
Evaluation, guardrails, on-prem constraints, and a working session to pick use cases and write the adoption plan.
IIT engineer with a Master's in the US, and a former data scientist who built ML systems in production. At Grokking he leads AI research and engineering — designing the document-extraction and evaluation pipelines behind agents in use at enterprises like NHPC, Indian Oil, and Maruti Suzuki. This course teaches the exact craft that work depends on.
Basic Python — functions, loops, dicts — is enough, and we include a short primer if you're rusty. Every lab also ships with a working template, so non-coders can complete all of them; you just need to be comfortable running a notebook.
Patterns first, framework second. LangGraph and CrewAI, tools and data over MCP, and Claude and GPT-class models plus one open model — so you learn the code is model-agnostic.
It's the hardest and most valuable part of enterprise AI: answering from real documents without hallucinating, every answer traceable to the exact page. It's also exactly what Grokking builds commercially, so you learn it from production experience.
Yes — the corporate workshop is a 2–3 day, on-site or private version run on your own documents. Teams that go through it often move straight into a production build with us.
Three working agents you built, a capstone you can demo, a one-page write-up for each, and a certificate of completion. Standout capstones get featured on Grokking's channels.
Yes, and it is included in the ₹8,999. We rewrite your resume around the agents you shipped, run mock interviews, rebuild your LinkedIn, clean up your GitHub, and draft your first posts with you. We help you show the work well — we don't place people or promise offers.
Starts 15 August 2026, runs six weeks, ₹8,999 with everything included. Tell us a little about you and we’ll send joining details.