PG Diploma in AI & Machine Learning

Explain a prediction and its mistakes

Choose this focus if you want to understand why a model predicts a number or assigns a category, and where its answers are unreliable.

Connect Python and data exploration to regression and classification. Compare predicted and observed values before trying a more complicated model.

Check what an AI answer is based on

Choose this focus if your interest is document questions, prompting and RAG. A fluent answer is only one part of the result: the supporting passage matters too.

Connect prompting, embeddings and retrieval to answer evaluation. Keep the question, retrieved passage and answer together so another person can inspect the result.

Follow a reading or transaction through a system

Choose this focus if you want to connect software concepts with sensor readings or the state of a smart contract.

Use IoT architecture and communication topics for the sensor path, and Solidity, transactions and contract testing for the ledger path. Treat these as two separate exercises with separate checks.

Practice brief

A housing prediction with an error log

Try this

Use the synthetic housing dataset to explore area, bedrooms and age before trying a simple regression. Keep a set of rows aside to check predictions.

Keep this output

A notebook with the input columns, a predicted-versus-observed table, an error plot and short notes about the largest misses.

Check and explain

Keep price_lakh as the value to predict, not an input. Compare the model with a simple baseline on the same held-out rows, and explain which kinds of houses have larger errors. The teaching prices are not a property valuation.

Syllabus connection

• Python
• Data Science using Python
• Machine Learning

Inspect the synthetic housing CSV

Practice brief

A document answer with a source trail

Try this

Use a short document you are allowed to share. Prepare questions whose answers are in it and one question the document does not answer. Sketch or practise a retrieval-and-answer workflow.

Keep this output

An answer-review sheet containing each question, the retrieved passage, the answer and your judgement of whether the passage supports it.

Check and explain

Change one fact in the source and compare the next answer. For the unsupported question, check whether the response clearly identifies the missing information instead of inventing it.

Syllabus connection

• Generative AI

Practice brief

Two traces: sensor reading and contract state

Try this

Sketch a sensor-to-gateway-to-MQTT path with example readings. Separately, sketch a small Solidity record and two test transactions that change its state.

Keep this output

An IoT flow diagram and reading log, plus a contract state table showing before, transaction and after. Start with illustrative values when planning the exercise.

Check and explain

For IoT, mark units, timestamps and how a missing reading would appear. For the contract, show which transaction should succeed or fail and why. A working trace is the starting point for explanation; it does not establish production security.

Syllabus connection

• IoT
• Blockchain