Fundamentals of Generative AI
Duration
3 Days
Level
Intermediate to Advanced Level
Design and Tailor this course
As per your team needs
Overview
This course is an interactive, hands-on training designed to provide participants with a strong foundation in Generative AI concepts, architectures, and applications. The course blends theoretical understanding with practical implementation, enabling learners to design, build, and experiment with modern generative models across multiple data modalities.
Participants will gain exposure to key generative techniques used for text, image, and embedding generation, along with the underlying deep learning principles that power them. The program emphasizes conceptual clarity, experimentation, and real-world applicability, making it suitable for practitioners transitioning into Generative AI.
Audience
- Software Developers
- Data Engineers and Data Analysts
- Machine Learning Practitioners
- AI Engineers
- Technical Product Owners
Prerequisites
Participants should have:
- Working knowledge of Python programming
- Basic understanding of machine learning concepts
- Familiarity with neural networks and deep learning fundamentals
Curriculum
Introduction to Generative AI
- What is Generative AI and how it differs from traditional AI
- Generative vs discriminative models
- Enterprise and real-world use cases
- Overview of common generative model families
Deep Learning Primer
- Neural network fundamentals
- Activation functions and loss functions
- Training workflows and optimization basics
- Common challenges in training deep learning models
Core Building Blocks of Generative Models
- Probability distributions and sampling
- Latent space intuition
- Representation learning concepts
Hands-On Lab:
- Implementing a simple generative model using a deep learning framework
Variational Autoencoders (VAEs)
- VAE architecture and intuition
- Encoder–decoder workflow
- Reconstruction vs generation
- Use cases and limitations
Hands-On Lab:
- Building a VAE for image generation and reconstruction
Generative Adversarial Networks (GANs)
- Generator and discriminator roles
- GAN training dynamics
- Common instability issues
- Synthetic data generation use cases
Hands-On Lab:
- Training a GAN to generate synthetic samples
Advanced Generative Models
- Autoregressive generative models
- Flow-based models (high-level overview)
- Strengths and trade-offs of advanced approaches
Hands-On Lab:
- Implementing an autoregressive model for generation tasks
Text Generation Techniques
- Sequence modeling concepts
- Introduction to transformer-based generation
- Controlling output quality and diversity
Hands-On Lab:
- Building a text generation pipeline
Embeddings & Representation Learning
- What are embeddings and why they matter
- Word, sentence, and document embeddings
- Semantic similarity and search use cases
Vector Databases & GenAI Workflows
- Role of vector databases in GenAI systems
- Storing and retrieving embeddings
- Supporting semantic search and retrieval pipelines
Capstone Project
- Designing a simple Generative AI solution
- Applying learned techniques to a real-world dataset
- Evaluating model outputs and behavior
Duration
3 Days
Level
Intermediate to Advanced Level
Design and Tailor this course
As per your team needs
