Build a ChatGPT Chatbot For Your Data
Join 100+ SaaS founders and leaders who have utilised AI to gain a competitive edge.
This course is no longer available.
Explore other coursesHosted by
Mayo Oshin, Harrison Chase, Anton Troynikov, and Jerry Liu
AI experts and Founders at LangChainAI, Chroma and LlamaIndex
Course overview
Unlock the Power of AI: Build Your MVP In 2 Weeks Or Less
You don't want to work with AI. You want AI to work for you.
Learning how to effectively utilise large language models is one of the fastest ways to build and accelerate your startup or career today.
An AI chatbot trained on your data that works could transform your business and work in impactful ways:
• From a SaaS founder struggling to build a well-functioning MVP to having users and investors beg to use your product and fund your idea.
• From an overwhelmed developer to effectively turning ideas in your head it into a working prototype.
• From an executive in an organisation spending a ton of money and time hiring and training analysts to becoming one of the few organisations in your industry that can analyse 1000s of pages in minutes instead of days.
Now, you could attempt to figure out a viable solution on your own via "free" content and trial and error. But this would likely waste valuable time and resources.
It'll cost you between $8k+ for your average developer and closer to $20k+ for more competent developers to build your MVP from scratch over several months without guaranteed results.
Alternatively, you could look for an "off-the-shelf" plug and play solution, but you'd lose the ability to have:
• 100% ownership of your IP.
• 100% privacy of your data.
• 100% flexibility to customise for your use case.
But there's a better way…
By the end of this program, you’ll be equipped with solid foundational knowledge and tools to build an MVP that can accurately 'chat' with custom data and delight users.
You'll avoid the wasted time and cost of getting started by using our proprietary done-for-you AI Chatbot Starter Kit that can be installed within a few minutes without losing your ownership, privacy, or flexibility — no programming experience required.
Here's what you'll get in the program:
• Self-paced workshops covering step-by-step visual guidance on how to build an AI-powered chatbot for your data — our low-code approach makes learning easy.
• Access to live (and pre-recorded) question-and-answer sessions with AI experts on optimal strategies and best practices.
• "100% done-for-you" AI Chatbot starter kit for your documents: Quickly set up a basic AI chatbot (including a UI) with minimal time and no programming experience to extend for your use case.
• Access to live office hours with Mayo to resolve any critical questions and challenges you're facing.
• A free 30-minute AI consultation with Mayo to successfully implement an AI strategy for your use case ($300 value).
• Fast-track assignments to help you complete your MVP chatbot faster.
• Curated resources, including recommended tools, visual guides, and templates to simplify and understand complex concepts.
• Private community access to other founders, entrepreneurs, and executives building and deploying their MVP chatbot in every use case imaginable.
• Lifetime access to course material.
• Bonus: Exclusive training on how to "chat" with large databases and CSV files, and more.
• Bonus: Access to surprise guest speakers, including alums and industry experts.
• Bonus: Showcase your MVP to 20,000+ potential users across Twitter, newsletter, and potentially YouTube.
Who is this course for
01
SaaS founders & product managers looking to build a successful AI MVP for their industry use case.
02
Executive leaders looking to design and apply an effective AI strategy to transform their organisation and gain a competitive edge.
03
Senior technicians and engineers looking for the best development practices and tools to efficiently build an AI solution.
What you'll learn
How to train a "ChatGPT" AI chatbot on your private documents
Learn how to use the latest GPT-3.5-turbo and GPT-4 models to integrate AI with your data.
How to "chat" with multiple large PDF files
You'll learn how to use AI to ask questions and get accurate answers from large PDF files (000's of pages), doc, txt, and other formats.
How to use AI without compromising privacy
Learn the best practices for data privacy, whilst benefiting from AI capabilities.
How to optimise "prompts" for better AI accuracy
Learn the best practices for prompt engineering for more accurate responses from the AI model.
How to reduce hallucination and optimise speed
Learn the latest strategies to reduce AI inaccurate results and improve the speed of results.
Best practices and trends from AI experts
Get valuable insights from leading AI experts on upcoming trends and strategies to optimise AI for your use case.
What’s included
Live sessions
Learn directly from your instructors in a real-time, interactive format.
Lifetime access
Go back to course content and recordings whenever you need to.
Community of peers
Stay accountable and share insights with like-minded professionals.
Certificate of completion
Share your new skills with your employer or on LinkedIn.
Maven Guarantee
This course is backed by the Maven Guarantee. Students are eligible for a full refund up until the halfway point of the course.
Course syllabus
6 live sessions • 32 lessons
Week 1
- Visit the OpenAI playground link here.
- "Log in" if you've already signed up. If not click "Sign Up" to create your account.
- Once you're logged in, click on the navigation bar header called "Playground".
- Learn how to use "LangChain": a popular AI framework for building chatbots
- Learn how to set up your development environment (prerequisite)
Nov
2
Introduction: How to Achieve Your Goals During the AI Chatbot Program
Introduction: How ChatGPT Really Works
A deep dive into how ChatGPT works "under the hood."
Before you watch
Timestamps
00:01 - Introduction & AI chatbot demos
05:17 - How ChatGPT really works (openai playground)
12:19 - Temperature (controlling randomness)
15:38 - Tokens
19:49 - Prompts (instructions)
23:50 - Chat Model & Chat "Memory" (System, User, Assistant)
33:34 - How to chat with your data
40:56 - Embeddings
47:45 - AI chatbot architecture
LangChain Fundamentals: How to ChatGPT Your Data
Note: You don't need to use the codebase mentioned in this video. You already have access the AI chatbot starter kit including up-to-date functionality inside the private slack community channel.
Once you've completed the steps in the prelude video, it's time to dive into how your AI Chatbot and LangChain (the framework used to build it), works under the hood.
Use the README.md file in your codebase folder to help provide guidance alongside the video
By the end of this workshop, your AI chatbot should be up and running in your local environment.
Timestamps for main video:
0:00 - Introduction & Demo
2:05 - Chatbot architecture
8:20 - LangChain explained
9:40 - How to "ingest" your data
27:17 - Ingestion script
31:00 - How to chat with your data
32:00 - Prompt templates
36:00 - LLM Chains
45:00 - Retrieval QA Chain
52:00 - Conversational QA Chain
55:00 - Frontend application
Week 2
- Node isn’t installed
- Node version is outdated (run: `npm install @latest -g`)
- Yarn isn’t installed on your machine (run: `npm install -g yarn`)
- .env file is missing.
- .env file contains the wrong credentials (or clashes with global environment variables).
- Missing card on OpenAI’s billing account: https://platform.openai.com/account/billing/payment-methods
- You’ve run out of OpenAI credits: https://platform.openai.com/account/usage
- API key is invalid.
- Invalid environment and index name in your .env file.
- Vector dimensions is not set to 1536.
- Pinecone namespace is not in lowercase.
- Attempting to query by integer. Pinecone stores vectors in uuid.
- Pinecone indexes of users on the Starter(free) plan are deleted after 7 days of inactivity. To prevent this, send an API request to Pinecone to reset the counter before 7 days.
- Pinecone index is corrupted. Retry from scratch with a new Pinecone project, index, and cloned repo.
- If the troubleshooting above doesn’t work, install from scratch and try again with another computer system.
Nov
7
Office Hour: Building MVP AI Applications with LangChainAI
Nov
9
🏫 Module 1 & 2 Live Coaching with Mayo
Troubleshooting: How to Solve Common Installation Problems For Your AI Chatbot
Here are common categories and problems that cause installation issues.
Installation
OpenAI
Pinecone
Resetting
If you're still struggling to set up your chatbot please reach out to technical support in the private slack channel.
How to Deploy Your AI Chatbot to Production
Prelude
To deploy your AI chatbot for live use, we will use Github and Vercel. If you haven't registered for either yet, use the links below:
If you're new to git and version control systems, here are some brief guides to help you understand the basics:
Vercel alternatives for deployment:
Common AI Chatbot problems and solutions:
Week 3
Nov
14
🤝Meet the expert - Performance Optimisation: How to Reduce AI Hallucinations and Improve Latency
Nov
16
🤝Meet the expert - Q&A: How to Build Advanced High Performance AI Applications Using LlamaIndex (w/ Jerry Liu)
Nov
17
Graduation Ceremony: Showcase Your AI Chatbot
What students are saying
Course schedule
4-6 hours per week
Tuesdays & Thursdays
12pm EST
Office hours and Q&A sessions to ask questions and receive guidance on building your AI chatbot for your use case.
"Automatically becoming that AI expert"
Testimonial: Senior Product Designer/Marketer, Damiano Redemagni
What people are saying
John Beradi
Nikodem Grzesiak
Karl Yeh
Cris Ippolite
Fady Bassil
Shawn Hamilton
Travis
Damiano Redemagni
Want to train your team? Email us to learn how.