Curriculum projects

The 12 portfolio projects in the AI Engineering Program

Each module from 2 to 13 ends in one project: the challenge, the data it uses and what you build. These are the projects of the curriculum, not stories of past learners. 13 modules, 346 guided hours and 12 portfolio projects (4 of them built on Indian data: Indian-language text, GST invoices (images and PDFs), public Indian government documents in English and Hindi, Indian documents (fields to extract)). Self-paced, online, in English. Modules unlock in order. Lesson 1.1 is free.

Module 2: Python for AI

Sales-analytics command-line tool

Challenge: Turn a folder of sales data into answers from the command line, the way a working tool does.
Data used: A sales data file (CSV).
What you build: A command-line tool with argparse, tests and exit codes.

Module 3: Mathematics & Statistics for AI

Math-to-ML notebook

Challenge: Train a model without a library hiding the maths.
Data used: A small dataset handled in NumPy.
What you build: A notebook that trains a model by hand, every step explained.

Module 4: Data, SQL & Machine Learning in Practice

Credit-risk or churn model

Challenge: Decide who is likely to default or churn, and where to set the threshold so mistakes cost least.
Data used: An Indian-style tabular dataset, queried with SQL.
What you build: A SQL feature pipeline, three models compared, a cost-based threshold and a model card.

Module 5: Deep Learning & Transformers from Scratch Indian data

Mini-GPT from scratch

Challenge: Make language models stop being a black box.
Data used: Indian-language text.
What you build: A small GPT-style transformer trained from scratch, plus a sentence-embedding search demo.

Module 6: LLM Application Engineering Indian data

GST invoice to data service

Challenge: Get clean data out of messy invoices automatically.
Data used: GST invoices (images and PDFs).
What you build: A FastAPI service that returns validated JSON, with a simple UI and the cost per document measured.

Module 7: AI-Native Engineering with Coding Agents

Merged open-source pull request

Challenge: Contribute to real software with a coding agent, safely.
Data used: An open-source code repository.
What you build: A reviewed pull request for a real feature or fix, covered by tests.

Module 8: RAG Engineering Indian data

Bilingual policy assistant

Challenge: Answer citizens' questions from official documents, with sources.
Data used: Public Indian government documents in English and Hindi.
What you build: A bilingual RAG assistant with citations and a retrieval-quality report.

Module 9: Evals, Guardrails & AI Security

Eval suite and red-team report

Challenge: Know when the AI app breaks, before users do.
Data used: Test sets and red-team prompts for the Module 8 assistant.
What you build: An evaluation suite and red-team report, wired into CI so a regression fails the build.

Module 10: Agents & MCP

Finance-operations agent

Challenge: Reconcile invoices against a ledger without letting the agent act unchecked.
Data used: Invoices and a ledger, through MCP tools.
What you build: A finance-operations agent that asks for approval before any write and logs every step.

Module 11: Fine-Tuning, Open Models & Inference Indian data

Fine-tuned extraction model

Challenge: Beat a prompt-only baseline with a small fine-tuned model.
Data used: Indian documents with fields to extract.
What you build: A fine-tuned open model, evaluated on a held-out set and published on Hugging Face.

Module 12: Production AI & LLMOps

Production deployment

Challenge: Run the Module 10 agent like a real product.
Data used: The Module 10 agent and its evals.
What you build: A containerised deployment with CI/CD, eval gates, monitoring and a runbook.

Module 13: Capstone & Career Launch

Client capstone and case study

Challenge: Solve a real problem end to end, for a small business, an NGO or an open dataset you choose.
Data used: Chosen by you with the client or from an open dataset.
What you build: A spec, the build, evals, a deployment, a demo video and a case study.

Your first lesson is on us.

AI vs ML vs DL vs GenAI: deep notes, a hands-on lab and a quiz. No card. No catch.

Start lesson 1 free