Skip to content

AI & Data · Intermediate

Generative AI & Document Assistants
One month. A practical start.

Explore language models, prompting and document retrieval, then build a small assistant that shows its sources and handles unanswered questions.

4 weeks
one focused course
20 days
2 guided hours per day
20 hrs
independent practice
1 project
to build and demonstrate

Your starting point

Is this course
right for your batch?

Students who already write basic Python and want an introduction to practical AI applications.

What students should know

  • Python functions, lists and dictionaries
  • Basic notebook use
  • A trainer-tested model runtime or shared inference endpoint

Tools & lab setup

  • Python notebooks
  • Hugging Face Transformers
  • Sentence Transformers
  • Gradio
  • Small public or synthetic document collection

Your one-month learning plan

A milestone
every week.

Each week has five teaching days: 10 guided hours and 5 hours of independent practice. The last week includes project review and the final demonstration.

Week 1 · Days 15

Models and careful prompting

  • Tokens, context and inference
  • Run a small pretrained model in a supplied notebook
  • Prompts with instructions and examples
  • Hallucinations, privacy and model licences
Weekly deliverable

A prompt comparison notebook with observed successes and failures.

Week 2 · Days 610

Find relevant information

  • Prepare a small approved document set
  • Split text into traceable chunks
  • Create embeddings and compare similarity
  • Compare keyword and semantic retrieval
Weekly deliverable

A document search prototype with source labels.

Week 3 · Days 1115

Connect retrieval and generation

  • Build prompts from retrieved passages
  • Attach source references
  • Handle insufficient evidence
  • Create a simple Gradio interface
Weekly deliverable

A local assistant that answers from a bounded document set.

Week 4 · Days 1620

Evaluate and improve

  • Create a 20-question evaluation set
  • Check factual support and retrieval misses
  • Try simple untrusted-document instructions
  • Record latency, limitations and demo results
Weekly deliverable

A documented assistant and evaluation report covering answerable and unanswerable questions.

Your final project

Campus handbook assistant

Build a prototype over a small public or synthetic campus handbook. Show supporting passages, flag missing information and evaluate at least 20 questions. Use a pretested small model or a trainer-provided endpoint; model training is outside this month.

How learning is assessed

Prompt experiments, retrieval checks, a capstone evaluation report, and an individual explanation of one failure case and its remedy.

Students should explain their own decisions, show their work and document limitations. Any certificate conditions are agreed with the college before the batch begins.

Keep learning

Official learning resources.

References for preparation and continued practice. These resources support the syllabus; their publishers do not endorse this training.

For departments & placement cells

Bring practical training
to your campus.

Share your students’ background, preferred course and dates. We’ll help shape a one-month batch around your college’s timetable and lab facilities.

Batch size, delivery arrangements, fees and any completion certificate requirements are agreed with your college before enrolment.

Tell us about your batch

We use these details to respond to your enquiry. Read our privacy policy.