Department of CSE, BUET

Home
------------------------------------------------------------

Application Open [Batch 33]

Admission deadline: (Batch 33)
2026-05-15

------------------------------------------------------------

Routine [Batch 33]
------------------------------------------------------------

Ongoing Courses (Batch 33)

------------------------------------------------------------
Applied Machine Learning
------------------------------------------------------------
Generative AI and Deep Learning
------------------------------------------------------------
Introduction to Python
------------------------------------------------------------

IAC Facility Booking
------------------------------------------------------------
About Us
------------------------------------------------------------
Contact
------------------------------------------------------------
Login

Course Detail:

Generative AI and Deep Learning

Course Title: Generative AI and Deep Learning

Introduction

The Generative AI and Deep Learning course at the Bangladesh-Korea Information Access Center (BK-IAC), Department of CSE, BUET, is designed for learners who want to develop a practical understanding of modern Generative AI. This course introduces the essential foundations of artificial intelligence, machine learning, deep learning, generative models, Transformers, Large Language Models (LLMs), prompt engineering, fine-tuning, and Retrieval-Augmented Generation (RAG).

The course follows a hands-on approach using Python-based tools and modern AI libraries. Participants will learn how generative AI models work, how they are applied in real-world systems, and how to build practical applications such as text generation systems, image generation models, chatbots, and document question-answering systems. The course also introduces responsible AI practices, including limitations, hallucination, bias, privacy, and ethical use of Generative AI.

Objectives

To introduce the fundamental concepts of artificial intelligence, machine learning, deep learning, and generative AI.
To provide a concise understanding of classical machine learning models and neural network foundations.
To develop hands-on skills in Python and PyTorch for building and experimenting with AI models.
To explain important deep learning architectures such as CNNs, RNNs, and Transformers.
To introduce major generative models, including Autoencoders, Variational Autoencoders, GANs, and diffusion-style generative systems.
To provide practical understanding of Large Language Models, including BERT, GPT-style models, LLaMA-style models, and chatbot systems.
To train participants in prompt engineering, basic fine-tuning, model adaptation, and Retrieval-Augmented Generation.
To apply Generative AI skills through practical assignments and a final case-study/project.
To create awareness of responsible AI practices, including hallucination, bias, privacy, copyright concerns, and ethical deployment.

Prerequisite

We expect participants to have prior programming knowledge, specifically in Python. Basic understanding of mathematics, including linear algebra and probability, will be helpful. Prior knowledge of machine learning is recommended, but the course will briefly review the essential concepts required for understanding Generative AI.

This course involves hands-on programming, model training, and project work. Enrolling without adequate Python knowledge may lead to significant difficulties in understanding the content and completing the assignments. This responsibility lies solely with the participant. We have also offered an "Introduction to Python" course (Course Details Link), which participants may complete prior to enrolling in this course.

Tentative Class Schedule

The course length will be 8 weeks with two classes in each week and 3 hours in each class. The tentative lecture plan of the course is as follows:

Class#	Content
1	Introduction to Artificial Intelligence, Machine Learning, and Generative AI: Concepts, Applications, Current Trends, and Responsible Use
2	Machine Learning Fundamentals: Classification, Regression, Training, Validation, Model Evaluation, and Overview of Classical ML Algorithms
3	Neural Network Foundations: Perceptron, Multilayer Networks, Activation Functions, Loss Functions, Backpropagation, and Gradient Descent
4	Introduction to PyTorch: Tensors, Autograd, Dataset and DataLoader, Model Definition, Training Loop, and Practical Neural Network Implementation
5	Deep Learning Architectures - I: Convolutional Neural Networks, Image Processing, Transfer Learning, and Practical CNN Applications
6	Deep Learning Architectures - II: Sequential Data, Word Embeddings, RNN, LSTM, GRU, and Introduction to Text Generation
7	Introduction to Generative Modeling: Discriminative vs. Generative Models, Latent Space, Sampling, Autoencoders, and Variational Autoencoders
8	Generative Adversarial Networks and Image Generation: GAN Framework, Generator, Discriminator, DCGAN, Conditional GAN, Style-Based Generation, and Assignment 1
9	Transformers - I: Attention Mechanism, Self-Attention, Query-Key-Value, Positional Encoding, and Transformer Architecture
10	Transformers - II: Encoder Models, Decoder Models, BERT, GPT-style Models, Tokenization, and Self-Supervised Pretraining
11	Large Language Models: Instruction-Tuned Models, ChatGPT-style Systems, LLaMA-style Models, Code Models, Multilingual LLMs, and Real-World Applications
12	Prompt Engineering: Zero-Shot and Few-Shot Prompting, Prompt Design Principles, Structured Prompts, Reasoning Prompts, and Prompt Evaluation
13	LLM Fine-Tuning and Adaptation: Transfer Learning, Domain-Specific Fine-Tuning, Parameter-Efficient Fine-Tuning, LoRA, Model Evaluation, and Assignment 2
14	Retrieval-Augmented Generation: Embeddings, Vector Databases, Document Chunking, Semantic Search, RAG Pipeline Design, and Hallucination Reduction
15	Generative AI Applications and Deployment: Chatbots, Document Assistants, Image Generation Tools, Model Compression, Quantization, On-Device AI, and Responsible Deployment
16	Final Project Evaluation: Presentation of Generative AI Case Studies, Project Demonstration, Discussion on Career Pathways, Research Directions, and Future Trends

Hands-On Projects

Text generation using RNN/LSTM or Transformer-based models.
Image generation using Autoencoders, Variational Autoencoders, or GANs.
Prompt engineering and evaluation using modern Large Language Models.
Fine-tuning or adapting a language model for a domain-specific task.
Building a chatbot or document question-answering system using Retrieval-Augmented Generation.
Designing a final Generative AI application and presenting the results.

Learning and Evaluation Method

Classes will be conducted in a multimedia-equipped environment or through an online interactive platform.
Expert faculty members from the field of artificial intelligence, machine learning, and deep learning will lead each class.
All instructors are faculty members or domain experts affiliated with the Department of CSE, BUET.
Participants will have access to necessary computing facilities for practical exercises, ensuring an interactive learning experience.
The course will include lectures, live coding sessions, lab exercises, assignments, case studies, and project work.
Participants will use Python, PyTorch, Scikit-Learn, Hugging Face libraries, and other relevant tools for practical implementation.
Evaluation will take place through assignments, hands-on projects, quizzes, and a final project presentation.
We will provide a certificate upon successfully passing this course.

Further Query

Email: iac@cse.buet.ac.bd
Phone: 9665650-80 Ext-6438
Mobile: 01670032959