Generative AI for Developers & Tech Leads (onsite) – 08 Feb 2024

Generative AI, when understood and applied correctly, will give you and your team superpowers.

”I’ve coded for 30+ years, but this tech has radically changed the way I build and design products! I never thought it would be possible to be this productive. And this is just the beginning.” – Henrik Kniberg.

The purpose of this 1-day course is to give you inspiration, knowledge, and tools to get started with AI-powered development in your organization. We’ll draw the big picture while also getting concrete with real-life examples (such as this) and hands-on coding exercises. We’ll mainly focus on GPT4 and the OpenAI API, but we’ll also look into other models. Bring your laptop, there will be some coding.

Note: this is an onsite course in Stockholm. Find remote versions here.

Key areas

• Big picture. What is Generative AI all about, beyond the hype? How does this impact software development in general?
• Personal productivity. How can Generative AI be used in your day-to-day work to greatly boost your personal productivity as developer or tech leader?
• Building AI-powered products. How can Generative AI techniques be used to turbocharge your products and services and give them capabilities that were previously impossible?

Highlights

• AI pair-programming patterns. Using tools like GPT4 as a coding buddy – patterns, tips, pitfalls. Integrating AI into your IDE.
• Prompt engineering, basic and advanced. We will deep-dive into prompt engineering, as it is a crucial skill for both personal usage (i.e. using chatgpt for coding and for other tasks), and developing AI-powered products (i.e. designing the prompts for your product to use).
• Embeddings & Vector DBs & Memory management. How to use techniques such as Vector Search and Retrieval Augmented Generation (RAG) to make your products more intelligent.
• Tools and function calling. How your AI code can break out of the AI sandbox with tools to edit files, integrate with third party services, etc.
• Automous Agents. How to build AI agents that act autonomously and don’t just wait for request/response.
• Limitations – How do deal with limitations like context length, data security, performance and cost.
• Ethics and safety – with technology as powerful as this, how do we avoid accidentally causing a disaster?
• Reflection and peer-coaching – discussion with your classmates, help figure out how to best apply this technlogy in your context

Prequisites and preparation

This course is aimed primarily at developers and tech leaders. You will need a working development environment (python or javascript/nodejs recommended). It’s OK if you are rusty and haven’t coded for a long time, with GPT 4 it is surprisingly easy to get up and running with only limited coding knowledge.

Preparation:

• ChatGpt 4 access. You will need a ChatGpt Plus account ($20 per month), or some other way to access the GPT 4 model. We will be using GPT 4 in our exercises.
• OpenAI account. You will need an OpenAI account, since we will use the OpenAI API in our coding exercise. It is pay-per-use and you get some starting credits, but you will probably need to add a few $ more. The cost is neglible for small-scale use (like in a course….).
• Bonus: Get OpenAI Hello World up and running. Follow the OpenAI Getting Started guide in python or nodejs and get GPT to say hello to you via code. If you run into trouble with this, don’t worry, we can help you during the course. But the course will be most valuable if you get this done before.