Empowering Educators in AI, ML, NLP & Generative AI
Duration – 5 Day- FDP
Objectives
To enable faculty members to gain hands-on
exposure to core AI/ML concepts, tools, and
implementation techniques, empowering them to
guide students effectively and integrate data
science modules in engineering curriculum and
research.
Tools & Platforms
• Python
• NumPy
• Pandas
Pre-requisites
• Basic knowledge of Python
• Interest in AI/ML and data handling
• Familiarity with math basics (like averages, graphs) is
helpful
Take away
• Strong foundation in AI/ML concepts and workflows
• Hands-on experience with Python, Scikit-learn, and
TensorFlow
• Ability to build and evaluate ML, DL, and NLP models
• Skills to work with real-world datasets
• Confidence to start projects and explore AI tools
independ
Day 2: LLM Internals, Prompt Engineering & ChatGPT
• Language Modelling & LLM Architecture
o Basics of Language Models (LMs)
o Fine-tuning Language Models: hands-on overview
o Pretraining vs. Fine-tuning: Use case comparisons
o Understanding Word Embeddings
o Transformers: How they work and architecture recap
• Prompt Engineering Deep Dive
o Building LLM solutions using prompts, embeddings, vector databases
o Types of prompting: Zero-shot, Chain of Thought, Tree of Thought
o Designing prompt templates
• ChatGPT as a Tool
o How ChatGPT works
o Hands-on exploration of ChatGPT for teaching/research use cases
o Introduction to other AI tools.
• Engagement Group activity: Create a Gen AI-based student assignment or classroom assistant
Day 3: Introduction to Generative AI & Project Lifecycle
• Understanding Generative AI
o What is Generative AI? How it works
o LLMs (Large Language Models) as creative thought partners
o Generative AI for writing, reading, summarizing, image generation
o What LLMs can and cannot do
• Prompting Fundamentals
o Tips and techniques for effective prompting
o Role of image and multimodal generation in Gen AI
• Project Lifecycle in Generative AI
o Using Gen AI in real applications
o Lifecycle: Ideation → Tool selection → Execution → Testing
o Cost considerations in AI model development
o Introduction to Retrieval Augmented Generation (RAG)
o Pretraining vs Fine-tuning of LLMs
o Instruction tuning and Reinforcement Learning from Human Feedback (RLHF)
o Tool use and intelligent agents
• Engagement Quiz / Reflection on academic use ca
Day 4: Gen AI Tools, LangChain & Application Development
• LLM Frameworks and Tools
o Chaining with LLMs: Key concepts and LangChain overview
o Multimodal AI applications with LangChain & OpenAI API
o Hands-on demo: Building a question-answering system
o Metrics to evaluate LLM applications
o Limitations and challenges of LLM deployments
• Gen AI Web Apps with LlamaIndex
o What is RAG? Role in Gen AI apps
o Build a basic full-stack Gen AI application
o Agent-based query handling and production best practices
• Vector Databases
o Vector representations and embeddings
o Searching techniques: ANN, Sparse/Dense/Hybrid
o Use case: Multilingual search applications
• Engagement
• Quiz / Peer feedback on prototype ideas
Day 5: Open-Source Tools, Azure OpenAI & Responsible AI
• Hugging Face & Open-Source Models
o Model selection for NLP and Vision tasks
o Applications: Translation, Summarization, Audio Classification, TTS, Object Detection, etc.
o Multimodal AI: Image captioning, retrieval, VQA, zero-shot learning
o Model deployment best practices
• Azure OpenAI Integration
o Building NLP solutions with Azure
o Prompt engineering on Azure platform
o Generating text/code/images with Azure OpenAI APIs
o Implementing RAG on Azure
• LLMOps Overview
o Data preparation and orchestration pipelines
o Safety and governance in AI applications
• Wrap-up & Reflection
o Showcase of top Gen AI use cases in industry
o Next steps: Certifications and learning paths
• Feedback, open Q&A, and closing
