Sign in Get started

Short course · Beginner to intermediate

GenAI for freshers: what placement panels actually ask

Embeddings, RAG and LLM internals — explained at the depth an interview panel expects, not at the depth a research paper expects.

6 lessons 53 min of reading Free · no account Practice set at the end
Start lesson 01 Take the practice set

Why this course exists

Almost every GenAI tutorial online is written for working engineers shipping a product. An interview panel is not asking you to ship a product — it is asking whether you understand what you built and can defend it for four minutes. That is a different depth, and nobody teaches it.

What you need first
  • Comfortable reading Python
  • No machine learning background needed
Who it is for

Pre-final and final year engineering students

What you will be able to do afterwards

  • Explain what an embedding is, and why cosine similarity is the right question to ask of one
  • Draw a RAG pipeline on a whiteboard and defend every arrow in it
  • Answer "how does an LLM actually work?" without either hand-waving or reciting the transformer paper
  • Say honestly where your own project would break — the answer panels are really testing for
  • Recognise the four follow-up questions that catch out students who only followed a tutorial

The 6 lessons

Read in order, or jump to what you need.

01 What panels actually ask about AI 7 min The four question shapes that come up, why "I used the OpenAI API" is a losing answer, and the standard you are being measured against. 02 How an LLM actually works 9 min Next-token prediction, why that explains hallucination, and a four-sentence answer that satisfies a panel without reciting the transformer paper. 03 Embeddings and vectors 9 min What an embedding really is, why cosine similarity and not distance, and the two follow-ups that catch out students who only used a vector database. 04 RAG — the diagram you will be asked to draw 10 min The two phases, every arrow in the pipeline, the three places it breaks, and why "RAG stops hallucination" is the wrong claim. 05 Prompting without the folklore 7 min What actually changes an output, what is cargo cult, and why the most valuable prompting skill in a real system is handling the output you get back. 06 Build a project that survives a technical round 11 min A scoped RAG project you can finish in a weekend, what to measure so you have numbers to quote, and the four follow-up questions it will attract. ✓ GenAI screening set 12 min 15 questions, marked instantly with an explanation for every option.
Start lesson 01

Also free

The other courses on this site

Free to learn · paid to be measured

Read the courses for nothing. Pay when you want to be marked.

Every lesson on this site is free and needs no account. What a subscription buys is the graded practice, the full-length mocks, the spoken AI interviews and a readiness score you can act on.

Start a free course See pricing