AI Engineering

This program on modern Artificial Intelligence systems, covers from the foundations of Transformer architectures to the development of advanced AI agents and autonomous systems. Learn how Large Language Models (LLMs) such as GPT, LLaMA, Mistral, and T5 are designed, trained, fine-tuned, and deployed for real-world applications. The curriculum provides hands-on experience in building intelligent applications using Retrieval-Augmented Generation (RAG), vector databases, prompt engineering, and multimodal AI technologies.

Students will develop the ability to create AI-powered chatbots, knowledge assistants, content generation systems, coding assistants, and enterprise AI solutions using industry-standard frameworks such as Hugging Face, LangChain, LlamaIndex, Ollama, ChromaDB, and FAISS. They will also learn to design, implement, and manage autonomous AI agents capable of reasoning, planning, memory management, tool usage, and decision-making.

The program further equips learners with skills in developing multi-agent systems, workflow automation, and no-code/low-code AI applications using platforms such as n8n, CrewAI, AutoGen, LangGraph, OpenAI Agents SDK, and Microsoft Copilot Studio. Through project-based learning, students gain practical expertise in deploying, monitoring, evaluating, and governing AI systems while addressing critical considerations such as ethics, safety, reliability, and scalability.

Upon completion, students will be capable of building next-generation intelligent applications and autonomous systems, preparing them for roles such as AI Engineer, Generative AI Engineer, LLM Engineer, AI Agent Developer, Machine Learning Engineer, AI Solutions Architect, and AI Product Developer.

Key skills you gain

Transformer Architecture: Self-attention mechanisms, encoder-decoder design, positional encoding, and multi-head attention for sequence modeling.
Large Language Models (LLMs): Pre-training, fine-tuning, prompt engineering, and deployment of large-scale language models such as GPT and BERT.
Generative AI: Principles of text, image, and multimodal content generation using diffusion models, GANs, and autoregressive architectures.
AI Agents: Design and implementation of goal-driven agents with tool use, memory, and reasoning capabilities in dynamic environments.
Agentic AI & Autonomous Systems: Multi-agent orchestration, autonomous decision-making pipelines, and real-world deployment of self-directed AI workflows.

CURRICULUM

Transformer Architecture

Learn the foundational architecture behind modern AI systems, where self-attention mechanisms enable models to process and relate information across entire sequences simultaneously. Students gain hands-on experience with encoder-decoder structures, multi-head attention, positional encoding, and feed-forward networks, along with model building and performance tuning using industry-standard deep learning frameworks such as PyTorch and TensorFlow.

Topics Covered

Foundations & Attention Mechanisms:	Limitations of RNNs and LSTMs, Positional Embeddings, Self-Attention Mechanism, Building a single attention head in PyTorch, Multi-Head Attention, Residual Connections & LayerNorm, Stacking into a Full Transformer Block.
Transformer Architecture Design:	Encoder Architecture, Decoder Architecture, Encoder-Decoder Transformers.
Modern Transformer Models:	BERT, GPT, T5, LLaMA, Mistral Architectures.
Applied Transformers & Fine-Tuning:	Hugging Face Transformers, Fine-Tuning Transformers, Transformer-based NLP and Vision Models.

Libraries & Frameworks:

PyTorch
TensorFlow / Keras
Hugging Face Transformers
PEFT & LoRA
LangChain / LlamaIndex

Large Language Models (LLMs)

Learn the principles and inner workings of Large Language Models, where billions of parameters are trained on massive text corpora to understand, generate, and reason across diverse language tasks. Students gain hands-on experience with pre-training strategies, prompt engineering, retrieval-augmented generation, and model fine-tuning, along with building and evaluating LLM-powered applications using industry-standard platforms such as Hugging Face, OpenAI API, and LangChain.

Topics Covered

LLM Fundamentals:	LLM Architecture, Pre-Training, Fine-Tuning.
Knowledge Augmentation & Reliability:	Retrieval-Augmented Generation (RAG), Vector Databases, Hallucination Handling.
Open-Source LLM Deployment:	Open-Source LLM Deployment, Ollama, LangChain, LlamaIndex.
LLM-Powered Application Development:	Chatbots, PDF Knowledge Assistant, Institutional AI Assistants, ChromaDB, FAISS

Libraries & Frameworks:

HuggingFace Transformers
LangChain
Ollama
LlamaIndex
ChromaDB / FAISS

Training a Tiny LLM

Learn the end-to-end process of training a compact Large Language Model from scratch, where raw text data is transformed into a fully functional language model through tokenization, embedding, and iterative gradient-based optimization. Students gain hands-on experience with dataset preparation, vocabulary building, loss computation, and training loop design, along with building and evaluating a fully operational tiny LLM using industry-standard deep learning frameworks such as PyTorch and Hugging Face Transformers.

Topics Covered

Tokenization & Data Preparation:	Byte-Level Tokenization, Dataset Batching & Shifting for Next-Token Prediction.
Loss Computation & Training:	Cross-Entropy Loss & Label Shifting, Training Loop from Scratch (No Trainer API).
Text Sampling & Generation:	Sampling: Temperature, Top-k, Top-p.
Model Evaluation & Validation:	Evaluating Loss on Validation Set

Libraries & Frameworks:

PyTorch
HuggingFace Transformers
Hugging Face Datasets
Tokenizers
NumPy
MatplotLib

Generative AI

Learn the core principles of Generative AI, where models are trained to create original content across text, image, audio, and video modalities by learning the underlying patterns and distributions of large-scale datasets. Students gain hands-on experience with diffusion models, GANs, VAEs, and multimodal generation pipelines, along with building and deploying creative AI applications using industry-standard tools such as Hugging Face Diffusers, Stable Diffusion, and OpenAI APIs.

Topics Covered

Generative AI Foundations & Architectures:	Generative AI Architectures, GPT Models, Tokenization, Embeddings.
Prompt Engineering & Multimodal AI:	Prompt Engineering, Multimodal AI Systems, AI Image Generation.
Generative AI Models & Platforms:	ChatGPT, Claude, Gemini, Hugging Face, Perplexity.
Applied Generative AI Development:	AI-Assisted Coding, Content Generation, Intelligent AI Assistants

Libraries, Frameworks, Platforms & API:

HuggingFace Diffusers
Tokenizers
LangChain
Gradio
Streamlit
OpenAI API
Goolge Gemini API
Stable Diffusion / DALL.E

AI Agents

Learn the design and development of AI Agents, where intelligent systems are built to perceive their environment, reason over goals, and autonomously execute multi-step tasks using tools, memory, and planning capabilities. Students gain hands-on experience with agent architectures, tool integration, memory management, and ReAct-based reasoning frameworks, along with building and deploying goal-driven AI agents using industry-standard platforms such as LangChain, LangGraph, and AutoGen.

Topics Covered

Foundations of Intelligent Agents:	Introduction to Intelligent Agents, Agent Architectures, Environments, Perception, Reasoning, Planning, Decision-Making.
LLM-Powered Agent Development:	Autonomous and Multi-Agent Systems, LLM Integration, Tool Integration, Memory Management, Retrieval-Augmented Generation (RAG).
Agent Design & Real-World Applications:	Conversational Agents, Task-Oriented Agents, Autonomous Agents, Real-World Application Design and Implementation.
Evaluation, Ethics & Deployment:	Evaluation, Deployment, Monitoring, Ethics, Safety, Governance of AI Agent Systems

Libraries, Frameworks, Platforms & API:

LangChain
LangGraph
AutoGen
Crew AI
Hugging Face Transformers

Agentic AI & Autonomous Systems

Learn the advanced principles of Agentic AI and Autonomous Systems, where AI models are designed to operate independently, coordinate across multi-agent pipelines, and execute complex real-world workflows with minimal human intervention. Students gain hands-on experience with agent orchestration, autonomous decision-making, self-reflective reasoning, and human-in-the-loop design, along with building and deploying end-to-end agentic systems using industry-standard platforms such as LangGraph, CrewAI, and AutoGen.

Topics Covered

Autonomous Agent Foundations:	Autonomous AI Agents, Planning and Reasoning, Memory Systems, Human-in-the-Loop Systems.
Multi-Agent Collaboration & Workflow Automation:	Multi-Agent Collaboration, Workflow Automation, No-Code/Low-Code AI Agent Development using n8n.
Agentic AI Frameworks & Platforms:	CrewAI, AutoGen, LangGraph, OpenAI Agents SDK, Microsoft Copilot Studio.
Enterprise Agentic Applications:	AI Scheduling Assistants, Research Agents, Enterprise Automation Systems