Advanced outline

Specialise: build, evaluate, and ship, including Indian-language AI.

Outline only. No certificate and not a university programme. Use it to decide what to build or evaluate next: then open official docs on AI4Bharat, Bhashini, or the employer ATS.

Advanced

Specialise: build, evaluate, and ship, including Indian-language AI.

About 8–12 hours plus a small project

Who it is for. Engineers, researchers, professors, and advanced builders targeting ML Engineer, research, or applied-science roles in India.

What you will learn

  • Prompting vs retrieval vs fine-tuning: what each is for, and what Indian employers usually mean by those words.
  • RAG, agents, and evaluation at a conceptual level: how you know a system works before you demo it.
  • Indian-language AI: why AI4Bharat, Bhashini, Krutrim, and Sarvam exist, and how they differ from English-first stacks.
  • Safety, data residency, and DPDP-aware design: enough to talk to policy, security, and hiring managers without bluffing.

Practice on your own work

Not graded. Not a certificate. Use a real circular, invoice, or ticket from your job or study.

  1. Drill 1

    Prompt vs retrieval vs fine-tune

    Pick one India task (Hindi circular, support ticket, or policy FAQ). Write which of the three you would try first and what evidence would make you switch.

  2. Drill 2

    Evaluate before you demo

    List ten real examples and a pass/fail check. Run them. If you cannot score the run, you are not ready to demo to a hiring manager or a ministry stakeholder.

  3. Drill 3

    Name an Indic stack

    Compare AI4Bharat, Bhashini, Krutrim, and Sarvam in three lines: what each is for, and what you would verify on their official page before you depend on it.

Back to Learn