LLM Engineering - The Foundations

4.6

(13 ratings)

4 Weeks

Cohort-based Course

LLM Engineering has expanded into our AI Engineering Bootcamp! Visit the link for more info! maven.com/aimakerspace/ai-eng-bootcamp

LLM Engineering - The Foundations

4.6

(13 ratings)

4 Weeks

Cohort-based Course

LLM Engineering has expanded into our AI Engineering Bootcamp! Visit the link for more info! maven.com/aimakerspace/ai-eng-bootcamp

Hosted by

Dr. Greg Loughnane and Chris "The LLM Wizard 🪄" Alexiuk

LLM Educators @ AI Makerspace and @ FourthBrain (Funded by Andrew Ng's AI Fund)

Dr. Greg Loughnane and Chris "The LLM Wizard 🪄" Alexiuk

LLM Educators @ AI Makerspace and @ FourthBrain (Funded by Andrew Ng's AI Fund)

Course overview

Understand LLM training from first principles

Master Large Language Model architecture, pretraining, prompt engineering, fine-tuning, and alignment. From the Transformer to RLAIF.

Data science and engineering teams around the world are being asked to rapidly build LLM applications through prompt engineering and supervised fine-tuning. Some companies are even working to train their own proprietary domain-specific LLMs from scratch!

In order to build a "ChatGPT" for your customers or internal stakeholders, using your own proprietary data, you'll want to understand how GPT-style models are actually built, step-by-step.

From closed-source models like OpenAI's GPT-series, Google's PaLM models, Anthropic's Claude, others, to open-source models like LLaMA 2-70B, Mistral-7B, 01.AI's Yi-34B, Mosaic's MPT-30B, or Falcon-180B, these decoder-only architectures are at the core, made in the same way.

This course will provide you with the foundational concepts and code you need to demystify how these models are created, from soup to nuts and to actually get started training, fine-tuning, and aligning your own LLMs.

From there, it's up to you to make the business case, organize the data, and secure the compute to give your company and your career a competitive LLM advantage.

Who is this course for

Aspiring AI Engineers looking to explore new career opportunities in Generative AI

Data scientists and Machine Learning Engineers who want to train their own LLMs

Stakeholders interested in training and deploying proprietary LLMs and applications

Aspiring AI Engineers looking to explore new career opportunities in Generative AI

Data scientists and Machine Learning Engineers who want to train their own LLMs

Stakeholders interested in training and deploying proprietary LLMs and applications

What you’ll get out of this course

Understand Large Language Model transformer architectures and how they process text for next word prediction

Grok how base-model next word prediction is improved for chat models and aligned with humans through RL

Deep dive on quantization strategies for more efficiently leveraging and fine-tuning LLMs including LoRA/qLoRA

Learn the story of of OpenAI's Generative Pre-Trained Transformers models from GPT to GPT-4 Turbo

Understand how Meta's LLaMA 2 was trained and why it serves as a benchmark for open-source LLM developers

Train models yourself each class using the classic techniques of pretraining, fine-tuning, and reinforcement learning

Work with other talented builders to bring your very own custom GPT-style Large Language Model to life

Explore frontiers of LLM Engineering including Mixture of Experts (MOE) models and Small Language Models (SLMs)

This course includes

Interactive live sessions

Lifetime access to course materials

10 in-depth lessons

Direct access to instructor

14 projects to apply learnings

Guided feedback & reflection

Private community of peers

Course certificate upon completion

Maven Satisfaction Guarantee

This course is backed by Maven’s guarantee. You can receive a full refund within 14 days after the course ends, provided you meet the completion criteria in our refund policy.

Course syllabus

Expand all modules

Week 1
Mar 19—Mar 24
Week dates are set to instructor's time zone
Modules
Session 1: The Transformer
Session 1: The Transformer
1 Lesson • 2 Projects
Session 2: From GPT to GPT-2
Session 2: From GPT to GPT-2
1 Lesson • 1 Project
Week 2
Mar 25—Mar 31
Week dates are set to instructor's time zone
Modules
Session 3: From GPT-2 to GPT-3
Session 3: From GPT-2 to GPT-3
2 Lessons • 2 Projects
Session 4: Quantization and PEFT-LoRA
Session 4: Quantization and PEFT-LoRA
1 Lesson • 1 Project
Week 3
Apr 1—Apr 7
Week dates are set to instructor's time zone
Modules
Session 5: Alignment I: Prompt Engineering & Instruction Tuning
Session 5: Alignment I: Prompt Engineering & Instruction Tuning
1 Lesson • 1 Project
Session 6: Alignment II: Reinforcement Learning with Human Feedback (RLHF)
Session 6: Alignment II: Reinforcement Learning with Human Feedback (RLHF)
1 Lesson • 3 Projects
Week 4
Apr 8—Apr 11
Week dates are set to instructor's time zone
Modules
Session #7: Reinforcement Learning with AI Feedback (RLAIF)
Session #7: Reinforcement Learning with AI Feedback (RLAIF)
1 Lesson • 2 Projects
Session #8: Frontiers of LLM Engineering
Session #8: Frontiers of LLM Engineering
1 Lesson • 1 Project
Post-Course
Modules
From LLM Engineering to LLM Operations
From LLM Engineering to LLM Operations
1 Lesson
Demo Day
Demo Day
1 Project

4.6

(13 ratings)

What students are saying

Prerequisites

A strong background in fundamental Machine Learning and Deep Learning

Understanding supervised learning, unsupervised learning, and neural network architectures are required. Introductory NLP and CV knowledge is encouraged.

A ability to program in Python within a Jupyter Notebook Environment

Understand basic Python syntax and constructs. You should be comfortable training and evaluating simple ML & DL models using test, train, and dev sets.

Free resource

📅 Detailed Schedule!

Understand how everything comes together in the course to provide a holistic overview of the how LLMs are engineered.

Get all the details about the assignments, associated papers, and key concepts you'll learn!

Send me the deets ✌️

Meet your instructors

Dr. Greg Loughnane

Founder & CEO @ AI Makerspace

I've worked as an AI product manager, university professor, data science consultant, AI startup advisor, and ML researcher; TEDx & keynote speaker, lecturing since 2013.

From 2021-2023 I worked at FourthBrain (Backed by Andrew Ng's AI Fund) to build industry-leading online bootcamps in ML Engineering and ML Operations (MLOps):

🔗 Resources links: Deeplearning.ai demos, AI Makerspace demos, LinkedIn, Twitter, YouTube, Blog.

Chris "The LLM Wizard 🪄" Alexiuk

Co-Founder & CTO @ AI Makerspace

I'm currently working as the Founding Machine Learning Engineer at Ox - but in my off time you can find me creating content for Machine Learning: either for the AI Makerspace, FourthBrain, or my YouTube Channel!

My motto is "Build, build, build!", and I'm excited to get building with all of you!

🔗 Resources links: YouTube, LinkedIn, Twitter

Be the first to know about upcoming cohorts