LLM Engineering: Master AI & Large Language Models (LLMs)
Ditulis pada: October 04, 2024
LLM Engineering: Master AI & Large Language Models (LLMs) - Master Generative AI and Large Language Models (LLMs). Explore and deploy LLM applications, learn fundamental theory.
Preview this Course
What you'll learn
- Project 1: Make AI-powered brochure generator that scrapes and navigates company websites intelligently.
- Project 2: Build Multi-modal customer support agent for an airline with UI and function-calling.
- Project 3: Develop Tool that creates meeting minutes and action items from audio using both open- and closed-source models.
- Project 4: Make AI that converts Python code to optimized C++, boosting performance by 60,000x!
- Project 5: Build AI knowledge-worker using RAG to become an expert on all company-related matters.
- Project 6: Capstone Part A – Predict product prices from short descriptions using Frontier models.
- Project 7: Capstone Part B – Execute Fine-tuneŠ² open-source model to compete with Frontier in price prediction.
- Project 8: Capstone Part C – Build Autonomous multi agent system collaborating with models to spot deals and notify you of special bargains.
- Design and develop a full solution to a given business problem by selecting, training and applying LLMs
- Compare and contrast the latest techniques for improving the performance of your LLM solution, such as RAG, fine-tuning and agentic workflows
- Weigh up the leading 10 frontier and 10 open-source LLMs, and be able to select the best choice for a given task
- Solve problems by applying leading open-source platforms, frameworks and tools, including Hugging Face, Gradio and Weights & Biases
- State the common AI paradigms, and identify the types of business problems most suitable for each
- Define fundamental data science concepts around deep learning, including training vs inference, generalizing vs overfitting, and the key ideas behind the NN
- Describe core concepts such as Generative AI, LLMs and the Transformer Architecture, and discuss what can be achieved with state-of-the-art performance
- Explain how LLMs work in sufficient detail to be able to train and test them, apply them to new scenarios, and diagnose & fix common issues
- Implement LLM solutions in Python using frontier and open-source models with both APIs and direct inference
- Execute code to write documents, answer questions and generate images.
Description
Mastering Generative AI and LLMs: An 8-Week Hands-On Journey
Accelerate your career in AI with practical, real-world projects led by industry veteran Ed Donner. Build advanced Generative AI products, experiment with over 20 groundbreaking models, and master state-of-the-art techniques like RAG, QLoRA, and Agents.
What you’ll learn
• Build advanced Generative AI products using cutting-edge models and frameworks.
• Experiment with over 20 groundbreaking AI models, including Frontier and Open-Source models.
• Develop proficiency with platforms like HuggingFace, LangChain, and Gradio.
• Implement state-of-the-art techniques such as RAG (Retrieval-Augmented Generation), QLoRA fine-tuning, and Agents.
• Create real-world AI applications, including:
• A multi-modal customer support assistant that interacts with text, sound, and images.
• An AI knowledge worker that can answer any question about a company based on its shared drive.
• An AI programmer that optimizes software, achieving performance improvements of over 60,000 times.
• An ecommerce application that accurately predicts prices of unseen products.
• Transition from inference to training, fine-tuning both Frontier and Open-Source models.
• Deploy AI products to production with polished user interfaces and advanced capabilities.
• Level up your AI and LLM engineering skills to be at the forefront of the industry.
About the Instructor
I’m Ed Donner, an entrepreneur and leader in AI and technology with over 20 years of experience. I’ve co-founded and sold my own AI startup, started a second one, and led teams in top-tier financial institutions and startups around the world. I’m passionate about bringing others into this exciting field and helping them become experts at the forefront of the industry.
Projects:
Project 1: AI-powered brochure generator that scrapes and navigates company websites intelligently.
Project 2: Multi-modal customer support agent for an airline with UI and function-calling.
Project 3: Tool that creates meeting minutes and action items from audio using both open- and closed-source models.
Project 4: AI that converts Python code to optimized C++, boosting performance by 60,000x!
Project 5: AI knowledge-worker using RAG to become an expert on all company-related matters.
Project 6: Capstone Part A – Predict product prices from short descriptions using Frontier models.
Project 7: Capstone Part B – Fine-tuned open-source model to compete with Frontier in price prediction.
Project 8: Capstone Part C – Autonomous agent system collaborating with models to spot deals and notify you of special bargains.
Why This Course?
• Hands-On Learning: The best way to learn is by doing. You’ll engage in practical exercises, building real-world AI applications that deliver stunning results.
• Cutting-Edge Techniques: Stay ahead of the curve by learning the latest frameworks and techniques, including RAG, QLoRA, and Agents.
• Accessible Content: Designed for learners at all levels. Step-by-step instructions, practical exercises, cheat sheets, and plenty of resources are provided.
• No Advanced Math Required: The course focuses on practical application. No calculus or linear algebra is needed to master LLM engineering.
Course Structure
Week 1: Foundations and First Projects
• Dive into the fundamentals of Transformers.
• Experiment with six leading Frontier Models.
• Build your first business Gen AI product that scrapes the web, makes decisions, and creates formatted sales brochures.
Week 2: Frontier APIs and Customer Service Chatbots
• Explore Frontier APIs and interact with three leading models.
• Develop a customer service chatbot with a sharp UI that can interact with text, images, audio, and utilize tools or agents.
Week 3: Embracing Open-Source Models
• Discover the world of Open-Source models using HuggingFace.
• Tackle 10 common Gen AI use cases, from translation to image generation.
• Build a product to generate meeting minutes and action items from recordings.
Week 4: LLM Selection and Code Generation
• Understand the differences between LLMs and how to select the best one for your business tasks.
• Use LLMs to generate code and build a product that translates code from Python to C++, achieving performance improvements of over 60,000 times.
Week 5: Retrieval-Augmented Generation (RAG)
• Master RAG to improve the accuracy of your solutions.
• Become proficient with vector embeddings and explore vectors in popular open-source vector datastores.
• Build a full business solution similar to real products on the market today.
Week 6: Transitioning to Training
• Move from inference to training.
• Fine-tune a Frontier model to solve a real business problem.
• Build your own specialized model, marking a significant milestone in your AI journey.
Week 7: Advanced Training Techniques
• Dive into advanced training techniques like QLoRA fine-tuning.
• Train an open-source model to outperform Frontier models for specific tasks.
• Tackle challenging projects that push your skills to the next level.
Week 8: Deployment and Finalization
• Deploy your commercial product to production with a polished UI.
• Enhance capabilities using Agents.
• Deliver your first productionized, agentized, fine-tuned LLM model.
• Celebrate your mastery of AI and LLM engineering, ready for a new phase in your career.
Who this course is for:
- Aspiring AI engineers and data scientists eager to break into the field of Generative AI and LLMs.
- Professionals looking to upskill and stay competitive in the rapidly evolving AI landscape.
- Developers interested in building advanced AI applications with practical, hands-on experience.
