Learn AI Engineeringin 16 Weeks

Ramp up your understanding of LLMs, their learning methods, neural network basics, and scaling laws.

• Neural networks: Parameters, Backpropagation, Loss, Activation functions

• Supervised, Self-supervised, Contrastive Learning, Reinforcement Learning

• Scaling Laws in LLMs

• Coding Assignment: MNIST

Understand the fundamental building blocks of LLMs with tokenization, vectorization and attention.

• Tokenisation

• Vectorization

• Positional Encodings

• Attention Mechanism

Dive into LLM internals with QKV matrices, Cross Attention, Multi-head Attention, and Feed Forward Neural Networks.

• Q, K, V Matrices

• Cross Attention

• Multi-head Attention

• FFNNs

• Logits

Code the internals of a transformer, following the architecture from the research paper: Attention is All You Need.

• Causal masked attention

• Model Training

• Auto-regressive models

• Coding Assignment: Transformer Internals

Learn the end-to-end lifecycle of an LLM, from data processing to training and inference.

• End-to-end LLM lifecycle

• Pre-training

• Post-training

• Model evaluation

Learn how LLMs are quantized for fast processing, and how to fine-tune models to meet specific business requirements.

• KV cache

• Quantization - FP16, attention optimizations

• Fine Tuning - LoRA/QLoRA

• Dataset prep → training → evaluation

Learn chunking strategies, data ingestion, reranking, indexing, vector databases, and other techniques for retrieval augmented generation.

• RAG: chunking strategies, data ingestion, reranker, indexer

• Vector Embeddings, Vector Databases

• Search Algorithms: ANN algorithms (HNSW, IVF)

An interactive project where students learn to code a RAG-based application and learn best practices for AI safety.

• Reranking strategies, Query rewriting, HyDE

• Input and output guardrails

• Safety: Prompt injection, Intent classification

• Coding Assignment: Build a RAG chatbot using API calls

Learn what an Agent is, how they are different from plain LLMs, Tool Calling, ReAct pattern, and Agent Orchestration.

• LLM vs Agent vs Multiple Agents

• ReAct pattern

• Prompt Chaining, Orchestration, Routing

• Coding Assignment: Customer support agent

Code an AI Agent with MCP and memory, optimizing agentic flow.

• Context Engineering

• Memory in Agents

• MCP, A2A

• Multi-Agents

• Coding Assignment: MCP with memory and optimising agentic flow

Learn how Evals are used in production AI applications, and best practices for AI development.

• Evals: How to avoid hallucinations with Evals

• LLM as a Judge

• Tradeoffs and design decisions

• Fine-tuning vs Prompting vs RAG

• Project: Build your own LLM Judge

Learn what reasoning models are and how they are trained using RHLF and Chain of Thought.

• Reasoning models

• RLHF

• Chain of Thought

• How to prompt a reasoning model

Learn how multimodal models are trained with images and video, and the mechanism of diffusion-based models.

• GANs

• Multimodal models

• CLIP

• Video Models

Learn the internals of diffusion-based models and why they are overtaking simple transformer architectures for image and video generation.

• Diffusion architecture

• Text-based diffusion

• Image and video generation

• Diffusion models vs LLMs

Build your industry-relevant AI project in two-weeks, using the lessons from the cohort.

• Choose an industry-relevant idea

• Use learnings from previous lessons

• Project review with AI Engineers

• Project showcase

Recap your learnings and learn best practices for building an AI system.

• Short recap

• Best practices when working with LLMs

• Limitations of GenAI at work

• Transitioning to AI Engineering