Large language models, the essentials at a glance!

This course covers machine learning theory and practical examples based on <Machine Learning, Just the Essentials—Fast!> (Insight, 2025).

This course covers the GitHub notebooks and bonus content for <Building an LLM from Scratch> (Gilbut, 2025). GitHub: https://github.com/rickiepark/llm-from-scratch/ <Building an LLM from Scratch> is the Korean translation of the bestseller <Build a Large Langauge Model (from Scratch)> (Manning, 2024) written by Sebastian Raschka. Starting from scratch and building a complete version of OpenAI’s GPT-2 model, this book provides a way to learn how large language models work and how to use them.

You will learn about Transformers inside out by directly implementing everything from Multi-Head Attention to the Original Transformer model, BERT, the Encoder-Decoder-based MarianMT translation model, and the Vision Transformer in code.

You will learn about RAG, VectorDB, LangChain, OpenAI, sLLM, and fine-tuning, which encompass the most important AI-related technologies at this time.

You use ChatGPT, but haven’t you found it difficult to explain why it produces answers like these? “I know terms like RAG, agents, and fine-tuning… but it’s difficult to explain them precisely.” “When I hear LLM-related terminology, I’m at a loss for words.” “Explanations of concepts in AI meetings are always vague.” This course was created specifically for people like you. This is a theory course designed to help you understand LLMs not as ‘tools,’ but as ‘structures.’ Rather than teaching you how to use ChatGPT or Gemini, it gives you a framework for explaining why they work the way they do.

This course covers basic natural language processing techniques and various text tasks using deep learning.

From core theories of deep learning and CNN to various CNN model implementation methods, and practical deep learning development know-how through real-world problems, If you want to become a deep learning CNN technology expert based on Pytorch, join us in this lecture :)

This course will teach you the concept of the Google Gemini model and how to use the Gemini API, and create various AI applications using Streamlit.

How do RNNs process sequential information? We’re releasing the RNN section of TTT: To The Transformer for free. Gain an intuitive understanding of how recurrent neural networks work and build a foundation for studying attention and transformers.

What if AI could directly open a browser and write shell scripts to resolve server issues? Through practical automation know-how implemented with OpenClaw across Windows and Linux, we guide you on the shortest path to completing your own powerful remote development environment without complex configurations.

Master the complete lifecycle of building modern Generative AI applications using Large Language Models, Retrieval-Augmented Generation, and Agentic AI systems. Learn to design enterprise-grade AI solutions from prompt engineering to deployment, combining LLMs with vector databases and external knowledge sources for production-ready applications.

LLM and Agent, the fundamental technologies behind AI like ChatGPT. LLM is currently the technology that has come closest to AGI. In this lecture, we will explore LLM and Agent technologies from a layperson's perspective.

We explore how to create an AI judge by combining knowledge graphs, reasoning systems, and various agent frameworks such as CrewAI, LangChain, and AutoGen, which enable the implementation of LLM-based multi-agent technology.

From perceptrons to backpropagation, systematically learn the core mathematical theories of neural networks and implement them directly in R. Build a solid foundation in deep learning—including activation functions, gradient descent, and multilayer neural networks—while developing the ability to solve classification problems using real-world datasets.

When you learn machine learning, you learn how to improve accuracy. But the moment you try to put that model onto an actual chip, questions no one has answered remain. How many multiplications does this model perform? How much does it need to remember, and for how long? Can the chip specifications you are targeting handle those requirements? Machine learning textbooks talk about accuracy, and semiconductor textbooks talk about circuits, but materials that bridge the gap between them are rare. I repeatedly encountered this gap while implementing RISC-V CPUs and NPUs on FPGAs. The model with the highest accuracy often turned out not to fit on the board. In this course, you will build six algorithms directly in Python, measuring the number of multiplications, comparisons, memory usage, and accuracy each time. Then you will convert those figures into the resources of an actual FPGA board. Saying that something requires a certain number of multiplications is not enough to gauge its size, but knowing that it uses a certain percentage of the board makes the decision clearer. After nine Labs, when the six models are brought together in a single table, an unexpected picture emerges. We spent nine lessons counting multiplications, but multiplications were not what held us back. After that, we reduce floating-point values to integers, manage bit widths, and even establish criteria for evaluating the circuit. You do not need an FPGA board. The course is conducted entirely in Python, and the board is used only as a reference for converting the measurements.

Putting a neural network on an FPGA—the phrase is familiar, but few have actually done it. If you want to stay a step ahead in capstone projects, competitions, or graduation projects, you must be able to go beyond just training a model and actually build and prove how those computations run inside a real chip. This course bridges that exact gap between theory and physical hardware. You will complete a neural network accelerator (NPU) for MNIST digit recognition from start to finish—from reading specifications to Python golden models, Verilog RTL design, and FPGA verification. From Conv layers made with line buffers and parallel MACs, through ReLU, MaxPool, and Flatten, to fully connected layers and Argmax—you will integrate nine blocks into a single chip and prove for yourself that "the Python reference and the hardware output do not differ by even a single bit (Bit-True)." I have even included how to overcome those frustrating moments when the simulation runs fine but the results are subtly different. I will guide you through practical know-how that general deep learning textbooks don't cover, such as integer quantization, array axis reordering, and accumulator overflow prevention. I have distilled the trial and error I experienced over years of RISC-V CPU design and FPGA Bit-True verification into this course so that you don't have to repeat them. By the end of the class, you will have a completed hardware project that allows you to say, "I designed and verified it myself."

Starting with the question, "If I were on the Titanic, would I have survived?", this course completes a full-stack project where you develop an AI model to predict survival probabilities based on real data and deploy it as a web service. You will gain hands-on experience in the entire process of AI and web development, from deep learning modeling using PyTorch to building a backend server with FastAPI and implementing a user interface with Next.js.

Using AI has become easier, but model cards and papers are still difficult to read, even just a few lines. Although you can look up the meanings of individual words, it’s not clear how those words connect to one another. This course builds up around 160 essential terms that frequently appear in AI documents, in the easiest order to understand them, so that in each lesson you can accurately read a single sentence from an actual document.

We will explore the types of applications and web services that can be built using LLMs.